Programming in Python

Programming YottaDB in the Python language is provided through a wrapper for non-threaded Simple API functions via a Python C extension created using the Python C API. This extension provides a yottadb module that may be imported into Python application code to enable programmatic access to YottaDB from within a Python programming environment. Note that the Python wrapper includes YottaDB as a dependency and so must be installed on a system after YottaDB is installed.

Since Python is a dynamically typed, object oriented language whereas C is statically typed and lacks object orientation, YDBPython abstracts away all C-level constructs in favor of Pythonic representations. This approach spares Python developers from the labors of managing memory and dealing with other “C-isms”.

The Python wrapper provides two ways of calling YottaDB API functions:

  • Python functions that directly call YottaDB API functions, mapping one Python interface to each YottaDB API function

  • Methods on a YottaDB Key class provided in the yottadb Python module

Note that the YDBPython doesn’t include any threaded YottaDB C API functions. These omissions are due to Python’s lack of support for thread-level parallelism, which is in turn due to the constraints of the Python Global Interpreter Lock. Accordingly, users seeking concurrent computation when programming YottaDB from Python will need to use process-level parallelism via the multiprocessing library module. An example of such parallelization is given in YDBPython/tests/test_threenp1.py.

As a matter of vocabulary, note that Python class methods like __init__() and __iter__() are called “magic methods” in this document, though they are also sometimes called “dunder” methods.

Python Quick Start

The YDBPython wrapper requires a minimum YottaDB release of r1.30 and is tested with a minimum Python version of 3.8. Python 2 is not supported. If the Python packages on your operating system are older, and the Python wrapper does not work, please obtain and install a newer Python version.

This section assumes that YottaDB has already been installed. One way to install YottaDB is described in the Quick Start section. After completing step 2 of that guide, Installing YottaDB, follow the instructions below to download, install, and test the Python wrapper:

  1. Install prerequisites: python3-dev and libffi-dev packages: sudo apt install python3-dev libffi-dev

  2. Set YottaDB environment variables: source $(pkg-config --variable=prefix yottadb)/ydb_env_set

  3. Enter code directory cd YDBPython/

  4. Run setup.py to install:

    1. Setup environment:

      1. Set the $ydb_dist environment variable: export ydb_dist=/path/to/my/ydb/installation

      2. Inform the linker of the location of YDB libraries: export LD_LIBRARY_PATH=$ydb_dist

    2. Install:

      • Option 1: install in venv (preferred method)

        1. Install the python3-venv package: sudo apt install python3-venv

        2. Create venv: python3 -m venv .venv

        3. Activate venv: source .venv/bin/activate

        4. Install into venv: python setup.py install

      • Option 2: install to user

        1. This method requires setuptools: sudo apt install python3-setuptools

        2. Install for use by user: python3 setup.py install --user

      • Option 3: Install globally (not suggested)

        1. This method also requires setuptools: sudo apt install python3-setuptools

        2. Install package globally: sudo -E python3 setup.py install

In the above instructions, note that python3 command is used when using a global Python 3 installation, i.e. one installed for the current system using e.g. apt-get install. The python command is used when operating within an active virtual environment (“venv”) as described above. The reason for the discrepancy is that many systems map the python command to Python 2, and use python3 to call a Python 3 installation. Within a virtual environment, Python binary paths are remapped to allow the python command to reference Python 3. The same principle applies to the pip command, with pip3 referencing the Python 3 version of the pip command. pip references the Python 2 implementation unless called within a virtual environment, where pip is an alias for pip3.

To run the tests for the Python wrapper and validate it was built and installed correctly:

  1. Install pytest and psutil

    1. If pip for python3 is not installed do so: sudo apt install python3-pip

    2. Use pip to install pytest and psutil

      • Option 1: install into venv

        1. Activate venv if it is not already: source .venv/bin/activate

        2. Install: pip install pytest psutil

      • Option 2: install for user: pip3 install --user pytest

      • Option 3: install globally (not suggested): sudo pip3 install pytest

  2. Run the tests:

    • Option 1: using venv: python -m pytest

    • Option 2 or Option 3: using user or global installation: python3 -m pytest

There are a number of test programs in the YDBPython/tests directory that you can look at for examples of how to use the Python wrapper.

To write your own programs using the YDBPython wrapper, simply import the yottadb module into your Python program with import yottadb after installing it via one of the methods specified above.

If you would like to import the yottadb module in a location outside of the YDBPython repository, you may do the following:

  1. Import yottadb from an arbitrary directory:

    • Approach 1: using a local YDBPython repository, e.g. as built above:

      • Option 1: using venv: pip install --editable /path/to/YDBPython/directory

      • Option 2 or Option 3: using user or global installation: pip3 install --editable /path/to/YDBPython/directory

    • Approach 2: using the PyPi package:

      • Option 1: using venv: pip install yottadb

      • Option 2 or Option 3: using user or global installation: pip3 install yottadb

Note that if using a virtual environment (“venv”), you will need to activate it with source .venv/bin/activate before using YDBPython in each new terminal session, and not only at installation time.

Python Concepts

As the YottaDB wrapper is distributed as a Python package, function calls to YottaDB are prefixed in Python code with yottadb. (e.g., application code to call the get() function would be written yottadb.get(...)). Alternatively, users may instantiate a Key object and use the methods on that object to call YottaDB API functions, e.g.:

key = yottadb.Key("^myglobal")["sub1"]["sub2"]
key.get()

Python Exception Handling

The YottaDB C API has a comprehensive set of error return codes. Each error is comprised of a unique number and a mnemonic. Thus, for example, to return an error that a buffer allocated for a return value is not large enough, YottaDB uses the “INVSTRLEN” error code, which has the numeric value yottadb.YDB_ERR_INVSTRLEN. YottaDB attempts to maintain stability of the numeric values and mnemonics from release to release, to ensure applications remain compatible when the underlying YottaDB releases are upgraded.

In contrast, Python applications typically use exceptions to handle errors, rather than numeric codes as C does. To reconcile these two different error handling mechanisms, YDBPython uses a hybrid approach by implementing, with a few exceptions (no pun intended), a generic yottadb.YDBError exception class with a YDBError.code() method for accessing the error code of the underlying error indicated by YottaDB. Each yottadb.YDBError exception raised will include an error message describing the failure. The YDBError.code() method is provided as a convenience in cases where a human-readable error message is insufficient and code needs to differentiate handling for different error scenarios.

Below are examples illustrating how to handle exceptions both with and without using the YDBError.code() method:

try:
    yottadb.get(varname="^myglobal", subsarray=("sub1", "sub2"))
except YDBError:
    print("Generic case: handle any error issued by YottaDB")

try:
    yottadb.node_next(varname="^myglobal", subsarray=("sub1", "sub2"))
except YDBNodeEnd:
    print("Specific case: handle YDB_ERR_NODEEND differently")

There are, however, a few special exceptions in YDBPython that are used to signal events that are not necessarily errors, but may need special handling. These are distinguished by unique exception classes apart from yottadb.YDBError:

  • yottadb.YDBTimeoutError: Raised when a YDBPython function that includes a timeout limit has taken longer than the specified limit to complete execution, e.g. Python lock().

  • yottadb.YDBTPRollback: See Python tp() for more information.

  • yottadb.YDBTPRestart: See Python tp() for more information.

For example:

try:
    yottadb.tp(callback, args=(arg1,arg2,))
except yottadb.YDBTPRestart:
    return

The Python wrapper will also raise exceptions whenever it encounters its own errors, which may occur independently of any interactions with YottaDB itself, for example when incorrect Python types are passed as arguments to wrapper code. In such cases, YDBPython will raise either a YDBPythonError with a message describing the error, or else it will raise a built-in Python exception, e.g. ValueError. Python built-in exceptions are used whenever possible, with YDBPythonError being raised in a handful of unique scenarios not covered by built-in exceptions.

Note that though all YottaDB error codes are implemented as Python exceptions, not all of these exceptions are expected at the Python level since many YottaDB error codes represent C-level issues that Python users are not in a position to address. For instance, the aforementioned “INVSTRLEN” error pertains to a C buffer allocation size error and so is not meaningful to a user of the Python wrapper.

Given the nature of exception handling, there is no “success” exception when a YDBPython wrapper function succeeds. At the C level, the YDB_OK code is returned. At the Python level, on the other hand, a successful call simply returns a value, if any, and omits to raise an exception. Accordingly, if an exception is raised, the call was not successful.

Python Symbolic Constants

YottaDB symbolic constants are available in the YDBPython module, for example, yottadb.YDB_ERR_INVSTRLEN.

Python API

YottaDB global and local variable nodes may be represented in multiple ways within the YDBPython wrapper. First, YottaDB nodes may be represented as two-element native Python tuples with the variable name as the first element of the tuple and a tuple containing a set of subscripts as the second element. For example, ("mylocal", ("sub1", "sub2")) represents the YottaDB local variable node mylocal("sub1","sub2"). Similarly, YottaDB nodes may be represented by tuples, e.g.: ("^test3", ("sub1", "sub2")). Unsubscripted local or global variable nodes may be represented by simply omitting the subscripts from the tuple or function call, for example: ("mylocal",) or yottadb.get("mylocal").

The Python wrapper also provides a Key class for interacting with YottaDB nodes in an object-oriented fashion. Each Key represents a combination of a global or local variable name and zero or more subscripts. Operations on this node may be performed by instantiating a Key object representing that node’s variable name and subscript combination and calling the method corresponding to the desired YottaDB API function on that object. For example:

key = yottadb.Key("^myglobal")["sub1"]["sub2"]
key.set("myvalue")
key.get()  # Returns b"myvalue"

Note that yottadb.get() and some other functions return Python bytes objects instead of str objects. This is because YottaDB stores arbitrary binary data, which is not guaranteed to be UTF-8 encoded, as Python str objects are by default. Accordingly, returning bytes objects allows users to retrieve arbitrary binary data from YottaDB without getting a UnicodeEncodeError for binary data that is not UTF-8 formatted. When accepting data (or subscripts, etc.), on the other hand, YDBPython accepts both str and bytes objects.

New Key objects may be created from existing Key objects by specifying additional subscripts in brackets, e.g.:

key1 = yottadb.Key("mylocal")  # key1 represents YottaDB node: `mylocal`
key2 = key1["sub1"]["sub2"]  # key2 represents YottaDB node: `mylocal("sub1","sub2")`

Intrinsic special variables may be accessed in the same way as global or local variables, with the provision that no subscripts are specified within the node tuple, as such variables are not actual YottaDB nodes. For example:

print(yottadb.get(("$ZYRELEASE",)))  # Print the current YottaDB release information

The length of strings (values and subscripts) in YottaDB is variable, as is the number of subscripts a local or global variable can have. However, in the case of the Python wrapper, such considerations are handled within the wrapper itself such that users need not concern themselves with memory allocation or management. Rather, users may simply pass valid Python objects to the wrapper (i.e. str, bytes, or, when setting values, int objects), which will take care of any memory allocation and management as needed.

Python API Functions

Python Data Structures & Type Definitions

As noted above, Python and C have significantly different approaches to data structures and memory management. Consequently, the YDBPython wrapper has no data structures that map directly to any C-level structure. Rather, the Python wrapper provides a combination of native Python tuples and Key objects for interacting with the underlying YottaDB C API.

Thus only one custom type is provided by the yottadb Python module:

  • Key an object class for representing a YottaDB local, global, or intrinsic special variable providing methods by which to access wrapper functions

All memory is managed internally and implicitly either by the YottaDB wrapper code (and YottaDB itself, for its own operations) or else by the Python runtime. Accordingly, users need not concern themselves with memory management or C-level data structures.

Python Wrapper Functions

Python ci()

def ci(routine: AnyStr, args: Tuple[Any] = (), has_retval: bool = False) -> Any

As a wrapper for the C function , the ci() function is used to call M routines from Python, used when a single call to the function is anticipated. ci() supports both read-only and read-write parameters.

If the specified routine has a return value, the caller of ci() must specify this using the has_retval parameter. This instructs the wrapper to internally allocate space for a return value and correctly construct the call to the underlying ydb_ci() YottaDB Simple API call. When there is no return value, None will be returned.

If a return value is specified but has not been configured in the call-in descriptor file or vice-versa, a parameter mismatch situation is created. In the parameter mismatch case, the error returned will be arbitrary and so may be inconsistent across calls. Accordingly, it is recommended to always ensure that routine parameters and return types are correctly specified in the call-in descriptor file.

  • args refers to a list of 0 or more arguments passed to the called routine. Arguments must be passed as Python str, bytes, or int objects. When calling routines that accept 0 arguments, the args field can simply be omitted or an empty Tuple passed (the default). Any output arguments will be returned as a Python bytes object and can be subsequently cast to another Python type. The number of parameters possible is restricted to 34 (for 64-bit systems) or 33 (for 32-bit systems). If the maximum number of parameters is exceeded, a ValueError will be raised.

  • has_retval is set to False by default. Accordingly, if the given routine has a return value has_retval will need to explicitly be set to True.

For example, see the below setup for a sample HelloWorld2 routine.

First, the call-in descriptor entry included in a call-in table file, e.g. calltab.ci:

HelloWorld2 : ydb_string_t * entry^helloworld2(I:ydb_string_t *, IO:ydb_string_t *, I:ydb_string_t *)

The contents of the M routine referenced by calltab.ci above, i.e. helloworld2.m:

; Hello world routine driven from Python
entry(p1,p2,p3)
    if ("1"'=p1)!("24"'=p2)!("3"'=p3) write "FAIL: parameters not as expected" quit "PARM-FAIL"
    set p2a=p2
    set p2="1"
    quit p3_p2a_p1

The Python call-in to the HelloWorld2 routine:

print("Python: Invoking HelloWorld2")
try:
    print(yottadb.ci("HelloWorld2", ["1", "24", "3"], has_retval=True))
except Exception as e:
    print(e)

The HelloWorld2 program in the example returns a string containing the three parameters, "1", "24", and "3" concatenated together in reverse order: "3241". Note that has_retval is set to True to signal that a return value is expected.

Note that a call-in table is required when calling from Python into M. A call-in table can be specified at process startup with the environment variable ydb_ci or using the functions yottadb.open_ci_table and yottadb.switch_ci_table, e.g:

cur_handle = yottadb.open_ci_table(cur_dir + "/tests/calltab.ci")
yottadb.switch_ci_table(cur_handle)

If the underlying ydb_ci() call returns an error, the function raises an exception containing the error code and message.

Python cip()

def cip(routine: AnyStr, args: Tuple[Any] = (), has_retval: bool = False) -> Any

As a wrapper for the C function ydb_cip(), the cip() function is used to call M routines from Python, used when repeated calls to the function are anticipated. Performance is slightly improved using cip() in such cases since this function saves a hash table lookup compared to ci(). cip() supports both read-only and read-write parameters.

If the specified routine has a return value, the caller of cip() must specify this using the has_retval parameter. This instructs the wrapper to internally allocate space for a return value and correctly construct the call to the underlying ydb_ci() YottaDB Simple API call. When there is no return value, None will be returned.

If a return value is specified but has not been configured in the call-in descriptor file or vice-versa, a parameter mismatch situation is created.

  • args refers to a list of 0 or more arguments passed to the called routine. Arguments must be passed as Python str, bytes, or int objects. When calling routines that accept 0 arguments, the args field can simply be omitted or an empty Tuple passed (the default). Any output arguments will be returned as a Python bytes object and can be subsequently cast to another Python type. The number of parameters possible is restricted to 34 (for 64-bit systems) or 33 (for 32-bit systems). If the maximum number of parameters is exceeded, a ValueError will be raised.

  • has_retval is set to False by default. Accordingly, if the given routine has a return value has_retval will need to explicitly be set to True.

For example, see the below setup for a sample HelloWorld3 routine.

First, the call-in descriptor entry included in a call-in table file, e.g. calltab.ci:

HelloWorld3 : ydb_string_t * entry^helloworld3(I:ydb_string_t *, IO:ydb_string_t *, I:ydb_string_t *)

The contents of the M routine referenced by calltab.ci above, i.e. helloworld3.m:

; Hello world routine driven from Python
entry(p1,p2,p3)
    if ("1"'=p1)!("17"'=p2)!("3"'=p3) write "FAIL: parameters not as expected" quit "PARM-FAIL"
    set p2a=p2
    set p2="1"
    quit p3_p2a_p1

The Python call-in to the HelloWorld3 routine:

print("Python: Invoking HelloWorld3")
try:
    print(yottadb.cip("HelloWorld3", ["1", "17", "3"], has_retval=True))
except Exception as e:
    print(e)

The HelloWorld3 program in the example returns a string containing the three parameters, "1", "24", and "3" concatenated together in reverse order: "3241". Note that has_retval is set to True to signal that a return value is expected.

Note that a call-in table is required when calling from Python into M. Additionally, any M routines that the call-in uses must be in a path referenced by the ydb_routines environment variable.

A call-in table can be specified at process startup with the environment variable ydb_ci or using the functions yottadb.open_ci_table and yottadb.switch_ci_table, e.g:

cur_handle = yottadb.open_ci_table(os.getcwd() + "/tests/calltab.ci")
yottadb.switch_ci_table(cur_handle)

If the underlying ydb_cip() call returns an error, the function raises an exception containing the error code and message.

Python data()

def data(varname: AnyStr, subsarray: Tuple[AnyStr] = ()) -> int

As a wrapper for the C function ydb_data_s() / ydb_data_st(), data() returns an integer value of 0, 1, 10, or 11 for the specified local or global variable node indicating what data may or may not be stored on or under that node. The meaning of these values is as follows:

  • 0: There is neither a value nor a subtree, i.e., the node is undefined

  • 1: There is a value, but no subtree

  • 10: There is no value, but there is a subtree.

  • 11: There are both a value and a subtree.

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the variable name node should be referenced without any subscripts.

  • If the underlying ydb_data_s() / ydb_data_st() call returns an error, the function raises an exception containing the error code and message.

yottadb.set("mylocal", ("sub1", "sub2"), "test")
print(yottadb.data("mylocal", ("sub1", "sub2"))) # Prints 1
print(yottadb.data("mylocal", ("sub1",))) # Prints 10
print(yottadb.data("mylocal", ("sub1", "sub2", "sub3"))) # Prints 0

yottadb.set("mylocal", ("sub1", "sub2", "sub3"), "test2")
print(yottadb.data("mylocal", ("sub1", "sub2", "sub3"))) # Prints 1
print(yottadb.data("mylocal", ("sub1", "sub2"))) # Prints 11

Python delete_node()

def delete_node(varname: AnyStr, subsarray: Tuple[AnyStr] = ()) -> None

As a wrapper for the C function ydb_delete_s() / ydb_delete_st(), delete_node() deletes the value stored at the given local or global variable node, if any, but leaves any subtree intact.

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the variable name node should be referenced without any subscripts.

  • If the underlying ydb_delete_s() / ydb_delete_st() call returns an error, the function raises an exception containing the error code and message.

yottadb.set("mylocal", ("sub1",), "test1")
yottadb.set("mylocal", ("sub1", "sub2"), "test2")
print(yottadb.get("mylocal", ("sub1",))  # Prints b'test1'
print(yottadb.get("mylocal", ("sub1", "sub2"))  # Prints b'test2'
yottadb.delete_node("mylocal", ("sub1",))
print(yottadb.get("mylocal", ("sub1",))  # Prints None
print(yottadb.get("mylocal", ("sub1", "sub2"))  # Prints b'test2'

Python delete_tree()

def delete_tree(varname: AnyStr, subsarray: Tuple[AnyStr] = ()) -> None

As a wrapper for the C function ydb_delete_s() / ydb_delete_st(), delete_tree() deletes both the value and subtree, if any, of the given local or global variable node.

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the variable name node should be referenced without any subscripts.

  • If the underlying ydb_delete_s() / ydb_delete_st() call returns an error, the function raises an exception containing the error code and message.

print(yottadb.data("mylocal", ("sub1", "sub2"))) # Prints 0
yottadb.set("mylocal", ("sub1", "sub2"), "test")
print(yottadb.data("mylocal", ("sub1", "sub2"))) # Prints 1
print(yottadb.data("mylocal", ("sub1",))) # Prints 10
yottadb.delete_tree("mylocal", ("sub1",))
print(yottadb.data("mylocal", ("sub1", "sub2"))) # Prints 0
print(yottadb.data("mylocal", ("sub1",))) # Prints 0

Python get()

def get(varname: AnyStr, subsarray: Tuple[AnyStr] = ()) -> Optional[bytes]

As a wrapper for the C function ydb_get_s() / ydb_get_st(), get() returns the value at the referenced global or local variable node, or intrinsic special variable.

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the variable name node should be referenced without any subscripts.

  • If the underlying ydb_get_s() / ydb_get_st() call returns an error of GVUNDEF or LVUNDEF, the function returns a value of None and does not raise an exception.

  • If the underlying ydb_get_s() / ydb_get_st() call returns an error other than GVUNDEF or LVUNDEF, the function raises an exception containing the error code and message.

  • Otherwise, it returns the value at the node.

print(yottadb.get("mylocal", ("sub1", "sub2"))  # Prints None
yottadb.set("mylocal", ("sub1", "sub2"), "test")
print(yottadb.get("mylocal", ("sub1", "sub2"))  # Prints b'test'

Python incr()

def incr(varname: AnyStr, subsarray: Tuple[AnyStr] = (), increment: Union[int, float, str, bytes] = "1") -> bytes

As a wrapper for the C function ydb_incr_s() / ydb_incr_st(), incr() atomically increments the referenced global or local variable node by the value of increment, with the result stored in the node and returned by the function. The value of the unit of incrementation may be passed as either a Python str or int object.

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the variable name node should be referenced without any subscripts.

  • If a value for the increment parameter is omitted, the default increment is 1.

  • If the underlying ydb_incr_s() / ydb_incr_st() call returns an error, the function raises an exception containing the error code and message.

print(yottadb.get("mylocal", ("sub1", "sub2"))) # Prints None
print(yottadb.incr("mylocal", ("sub1", "sub2"))) # Prints b'1'
print(yottadb.incr("mylocal", ("sub1", "sub2"))) # Prints b'2'

Python lock()

def lock(keys: Tuple[Tuple[Union[tuple, Optional["Key"]]]] = (), timeout_nsec: int = 0) -> None

As a wrapper for the C function ydb_lock_s() / ydb_lock_st(), lock() releases all lock resources currently held and then attempts to acquire the named lock resources referenced. If no lock resources are specified, it simply releases all lock resources currently held and returns.

Lock resources are specified by passing YottaDB keys as a tuple or list of Python tuple or yottadb.Key objects. Each tuple representing a key must be of the form (variable_name, (subscript1, subscript2, ...)), i.e. consist of two elements, a string representing a variable name and a tuple containing a series of strings representing subscripts, if any.

If lock resources are specified, upon return, the process will have acquired all of the named lock resources or none of the named lock resources.

  • If timeout_nsec exceeds yottadb.YDB_MAX_TIME_NSEC, a yottadb.YDBError exception will be raised where yottadb.YDB_ERR_TIME2LONG == YDBError.code()

  • If the lock resource names exceeds the maximum number supported (currently 11), the function raises a ValueError exception.

  • If keys is not a Tuple of tuples representing variable name and subscript pairs, or a series of yottadb.Key objects, then the function raises a TypeError exception.

  • If it is able to acquire the lock resource within timeout_nsec nanoseconds, it returns holding the lock, otherwise it raises a YDBTimeoutError exception. If timeout_nsec is zero, the function makes exactly one attempt to acquire the lock, which is the default behavior if a value for timeout_nsec is omitted.

  • If the underlying ydb_lock_s() / ydb_lock_st() call returns any other error, the function raises an exception containing the error code and message.

The following example provides a demonstration of basic locking operations. The example locks several keys, then attempts to increment the lock on each key by calling a separately defined lock_value() helper function as a separate Python process. Due to the initial locking of each key, each of these lock_value() fails with an exit code of 1. Next, all locks are released and a number of new lock_value() processes are spawned that again attempt to increment a lock on each key. Since all locks were previously released, these new attempts succeed with each process exiting with a 0 exit code.

import multiprocessing
import datetime

# Lock a value in the database
def lock_value(key: Union[yottadb.Key, tuple], interval: int = 2, timeout: int = 1):
    # Extract key information from key object to compose lock_incr()/lock_decr() calls
    if isinstance(key, yottadb.Key):
        varname = key.varname
        subsarray = key.subsarray
    else:
        varname = key[0]
        subsarray = key[1]
    if len(subsarray) == 0:
        subsarray = None

    # Attempt to increment lock on key
    has_lock = False
    try:
        yottadb.lock_incr(varname, subsarray, timeout_nsec=(timeout * 1_000_000_000))
        print("Lock Success")
        has_lock = True
    except yottadb.YDBTimeoutError:
        print("Lock Failed")
        sys.exit(1)
    except Exception as e:
        print(f"Lock Error: {repr(e)}")
        sys.exit(2)

    # Attempt to decrement lock on key, after a brief pause to ensure increment has taken effect
    if has_lock:
        time.sleep(interval)
        yottadb.lock_decr(varname, subsarray)
        if timeout != 0 or interval != 0:
            print("Lock Released")

    sys.exit(0)


t1 = yottadb.Key("^test1")
t2 = yottadb.Key("^test2")["sub1"]
t3 = yottadb.Key("^test3")["sub1"]["sub2"]
keys_to_lock = (t1, t2, t3)
# Attempt to get locks for keys t1,t2 and t3
yottadb.lock(keys=keys_to_lock, timeout_nsec=0)
# Attempt to increment/decrement locks
processes = []
for key in keys_to_lock:
    process = multiprocessing.Process(target=lock_value, args=(key,))
    process.start()
    processes.append(process)
for process in processes:
    process.join()
    print(process.exitcode)  # Prints 1
# Release all locks
yottadb.lock()
# Attempt to increment/decrement locks
processes = []
for key in keys_to_lock:
    process = multiprocessing.Process(target=lock_value, args=(key,))
    process.start()
    processes.append(process)
for process in processes:
    process.join()
    print(process.exitcode)  # Prints 0

Python lock_decr()

def lock_decr(varname: AnyStr, subsarray: Tuple[AnyStr] = ()) -> None

As a wrapper for the C function ydb_lock_decr_s() / ydb_lock_decr_st(), lock_decr() decrements the count of the lock name referenced, releasing it if the count goes to zero or ignoring the invocation if the process does not hold the lock.

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the variable name node should be referenced without any subscripts.

  • If the underlying ydb_lock_decr_s() / ydb_lock_decr_st() call returns an error, the function raises an exception containing the error code and message.

t1 = datetime.datetime.now()
yottadb.lock_incr("test2", ("sub1",))  # Increment lock on a local variable node, locking it
t2 = datetime.datetime.now()
time_elapse = t2.timestamp() - t1.timestamp()
print(time_elapse)  # Prints time elapsed, should be < 0.01
yottadb.lock_decr("test2", ("sub1",))  # Decrement lock on a local variable node, releasing it

Python lock_incr()

def lock_incr(varname: AnyStr, subsarray: Tuple[AnyStr] = (), timeout_nsec: int = 0) -> None

As a wrapper for the C function ydb_lock_incr_s() / ydb_lock_incr_st(), lock_incr() attempts to acquire the referenced lock resource name without releasing any locks the process already holds.

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the variable name node should be referenced without any subscripts.

  • If the process already holds the named lock resource, the function increments its count and returns.

  • If timeout_nsec exceeds yottadb.YDB_MAX_TIME_NSEC, a yottadb.YDBError exception will be raised where yottadb.YDB_ERR_TIME2LONG == YDBError.code()

  • If it is able to acquire the lock resource within timeout_nsec nanoseconds, it returns holding the lock, otherwise it raises a YDBTimeoutError exception. If timeout_nsec is zero, the function makes exactly one attempt to acquire the lock, which is the default behavior if timeout_nsec is omitted.

  • If the underlying ydb_lock_incr_s() / ydb_lock_incr_st() call returns any other error, the function raises an exception containing the error code and message.

t1 = datetime.datetime.now()
yottadb.lock_incr("test2", ("sub1",))  # Increment lock on a local variable node, locking it
t2 = datetime.datetime.now()
time_elapse = t2.timestamp() - t1.timestamp()
print(time_elapse)  # Prints time elapsed, should be < 0.01
yottadb.lock_decr("test2", ("sub1",))  # Decrement lock on a local variable node, releasing it

Python message()

def message(errnum: int) -> str

As a wrapper for the C function ydb_message() / ydb_message_t(), message() returns the text template for the error number specified by errnum. A negative error number is treated the same as its corresponding positive error number, such that yottadb.message(x) and yottadb.message(-x) produce the same output.

  • If errnum does not correspond to an error that YottaDB recognizes, a yottadb.YDBError exception will be raised where yottadb.YDB_ERR_UNKNOWNSYSERR == YDBError.code()

  • Otherwise, it returns the error message text template for the error number specified by errnum.

print(yottadb.message(-150375522))  # Prints '%YDB-E-INVSTRLEN, Invalid string length !UL: max !UL'

Python node_next()

def node_next(varname: AnyStr, subsarray: Tuple[AnyStr] = ()) -> Tuple[bytes, ...]

As a wrapper for the C function ydb_node_next_s() / ydb_node_next_st(), node_next() facilitates depth first traversal of a local or global variable tree.

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the variable name node should be referenced without any subscripts.

  • If there is a next node, it returns the subscripts of that next node as a tuple of Python bytes objects.

  • If there is no node following the specified node, a yottadb.YDBNodeEnd exception will be raised.

  • If the underlying ydb_node_next_s() / ydb_node_next_st() call returns any other error, the function raises an exception containing the error code and message.

# Initialize a test node and maintain full subscript list for later validation
subs = []
for i in range(1, 6):
    all_subs.append((b"sub" + bytes(str(i), encoding="utf-8")))
    yottadb.set("mylocal", subs, ("val" + str(i)))
# Begin iteration over subscripts of node
node_subs = ()
while True:
    try:
        node_subs = yottadb.node_next("mylocal", node_subs)
        print(node_subs)  # Prints (b'sub1',), (b'sub1', b'sub2'), etc. successively
    except yottadb.YDBNodeEnd:
        break

Python node_previous()

def node_previous(varname: AnyStr, subsarray: Tuple[AnyStr] = ()) -> Tuple[bytes, ...]

As a wrapper for the C function ydb_node_previous_s() / ydb_node_previous_st(), node_previous() facilitates reverse depth first traversal of a local or global variable tree.

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the variable name node should be referenced without any subscripts.

  • If there is a previous node, it returns the subscripts of that previous node as a tuple of Python bytes objects, or an empty tuple if that previous node is the root.

  • If there is no node preceding the specified node, a yottadb.YDBNodeEnd exception will be raised.

  • If the underlying ydb_node_previous_s() / ydb_node_previous_st() call returns any other error, the function raises an exception containing the error code and message.

# Initialize test node and maintain full subscript list for later validation
subs = []
for i in range(1, 6):
    all_subs.append((b"sub" + bytes(str(i), encoding="utf-8")))
    yottadb.set("mylocal", subs, ("val" + str(i)))
# Begin iteration over subscripts of node
node_subs = yottadb.node_previous("mylocal", subs)
print(node_subs)  # Prints (b'sub1', b'sub2', b'sub3', b'sub4')
while True:
    try:
        node_subs = yottadb.node_previous("mylocal", node_subs)
        print(node_subs)  # Prints (b'sub1', b'sub2', b'sub3'), (b'sub1', b'sub2'), and (b'sub1',), successively
    except yottadb.YDBNodeEnd as e:
        break

Python nodes()

def nodes(varname: bytes, subsarray: Tuple[bytes] = ()) -> NodesIter:

The nodes() function provides a convenient, Pythonic interface for iteratively performing depth-first traversals starting from the given YottaDB local or global variable node, as specified by the varname and subscripts arguments.

Specifically, nodes() returns a Python NodesIter iterator object that yields a List of subscripts representing the next node in the tree on each iteration, in accordance with the behavior for Python node_next().

Similarly, the reversed version of the returned NodesIter iterator will yield a List of subscripts representing the previous node in the tree on each iteration, in accordance with the behavior for Python node_previous().

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the variable name node should be referenced without any subscripts.

  • If there is a next node for a given iteration, the NodesIter iterator will return the subscripts of that next node as a tuple of Python bytes objects.

  • If this iterator is passed to the next() built-in function and there is no subscript following the subscript previously returned, a StopIteration exception will be raised.

  • If the underlying Python node_next() or Python node_previous() call returns any other error, the NodesIter iterator will raise an exception containing the error code and message.

# Create list of subscript arrays representing some database nodes
nodes = [
    (b"sub1",),
    (b"sub1", b"subsub1"),
    (b"sub1", b"subsub2"),
    (b"sub1", b"subsub3"),
    (b"sub2",),
    (b"sub2", b"subsub1"),
    (b"sub2", b"subsub2"),
    (b"sub2", b"subsub3"),
    (b"sub3",),
    (b"sub3", b"subsub1"),
    (b"sub3", b"subsub2"),
    (b"sub3", b"subsub3"),
]

# Set various nodes in the database
for node in nodes:
    yottadb.set("^myglobal", node, str(nodes.index(node)))

# Iterate over all nodes under a global variable
for node in yottadb.nodes("^myglobal"):
    # Prints: b'0', b'1', b'2', b'3', b'4', b'5', b'6', b'7', b'8', b'9', b'10', b'11'
    print(yottadb.get("^myglobal", node))

# Iterate over some nodes under a global variable, beginning at a
# subscript in the middle of the tree.
for node in yottadb.nodes("^myglobal", ("sub2",)):
    # b'5', b'6', b'7', b'8', b'9', b'10', b'11'
    print(yottadb.get("^myglobal", node))

# Iterate over all nodes under a global variable, in reverse order
for node in reversed(yottadb.nodes("^myglobal")):
    # b'11', b'10', b'9', b'8', b'7', b'6', b'5', b'4', b'3', b'2', b'1', b'0'
    print(yottadb.get("^myglobal", node))

# Iterate over some nodes under a global variable in reverse order,
# beginning at a subscript in the middle of the tree.
for node in reversed(yottadb.nodes("^myglobal", ("sub2",))):
    # b'7', b'6', b'5', b'4', b'3', b'2', b'1', b'0'
    print(yottadb.get("^myglobal", node))

Python open_ci_table()

def open_ci_table(filename: AnyStr) -> int

As a wrapper for the C function ydb_ci_tab_open(), the open_ci_table() function can be used to open an initial call-in table if the environment variable ydb_ci does not specify an M code call-in table at process startup. filename is the filename of a call-in table, and the function opens the file and initializes an internal structure representing the call-in table and returns an integer representing a handle for later reference to this call-in table.

After a successful call to open_ci_table(), YottaDB processes may then use the zroutines intrinsic special variable to locate M routines to execute. $zroutines is initialized at process startup from the ydb_routines environment variable.

If the underlying ydb_ci_tab_open() call returns an error, the function raises an exception containing the error code and message.

For an example of how to use open_ci_table, see the entry for Python ci() or Python cip().

Python release()

def release() -> str

Returns a string consisting of six space separated pieces to provide version information for the Python wrapper and underlying YottaDB release:

  • The first piece is always “pywr” to identify the Python wrapper.

  • The Python wrapper release number, which starts with “v” and is followed by three numbers separated by a period (“.”), e.g., “v0.90.0” mimicking Semantic Versioning. The first is a major release number, the second is a minor release number under the major release and the third is a patch level. Even minor and patch release numbers indicate formally released software. Odd minor release numbers indicate software builds from “in flight” code under development, between releases. Note that although they follow the same format, Python wrapper release numbers are different from the release numbers of the underlying YottaDB release as reported by $zyrelease.

  • The third through sixth pieces are $zyrelease from the underlying YottaDB release.

print(yottadb.release())  # Prints e.g. 'pywr v0.10.0 YottaDB r1.32 Linux x86_64'

Python set()

def set(varname: AnyStr, subsarray: Tuple[AnyStr] = (), value: AnyStr = "") -> None

As a wrapper for the C function ydb_set_s() / ydb_set_st(), set() updates the value at the referenced local or global variable node, or the intrinsic special variable to the value contained in the Python str or bytes object passed via the value parameter.

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the variable name node should be referenced without any subscripts.

  • If value is omitted, the node will be set to the empty string by default.

  • If the underlying ydb_set_s() / ydb_set_st() call returns an error, the function raises an exception containing the error code and message.

print(yottadb.get("mylocal", ("sub1", "sub2")))  # Prints None
yottadb.set("mylocal", ("sub1", "sub2"), "test")
print(yottadb.get("mylocal", ("sub1", "sub2")))  # Prints b'test'

Python str2zwr()

def str2zwr(string: AnyStr) -> bytes

As a wrapper for the C function ydb_str2zwr_s() / ydb_str2zwr_st(), str2zwr() provides the given string in Zwrite Format.

Note that the return value of this function is always a bytes object, reflecting the fact that YottaDB stores all values as binary data, such that a global or local variable node value is not guaranteed to be a valid UTF-8 string. Accordingly, the return value of this function is not guaranteed to be castable to a Python str object.

Further, note that the length of a string in Zwrite Format is always greater than or equal to the string in its original, unencoded format.

If the underlying ydb_str2zwr_s() / ydb_str2zwr_st() call returns an error, the function raises an exception containing the error code and message.

print(yottadb.str2zwr(b'X\x00ABC'))  # Prints b'"X"_$C(0)_"ABC"'

Python subscript_next()

def subscript_next(varname: AnyStr, subsarray: Tuple[AnyStr] = ()) -> bytes

As a wrapper for the C function ydb_subscript_next_s() / ydb_subscript_next_st(), subscript_next() facilitates breadth-first traversal of a local or global variable sub-tree. A node or subtree does not have to exist at the specified key.

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the subscript level is zero, and variable names should be iterated over instead of subscripts.

  • If there is a next subscript with a node and/or a subtree, this function returns the subscript at the level of the last subscript in subsarray

  • If there is no next node or subtree at that level of the subtree, a yottadb.YDBNodeEnd exception will be raised.

  • If the underlying ydb_subscript_next_s() / ydb_subscript_next_st() call returns any other error, the function raises an exception containing the error code and message.

In the special case where subsarray is empty, subscript_next() returns the name of the next global or local variable, and raises a yottadb.YDBNodeEnd exception if there is no global or local variable following varname.

yottadb.set("^myglobal", ("sub1", "sub2"), "val1")
yottadb.set("^myglobal", ("sub1", "sub3"), "val2")
yottadb.set("^myglobal", ("sub1", "sub4"), "val3")
yottadb.set("^myglobal", ("sub1", "sub5"), "val4")

# Get first subscript of the second subscript level
subscript = yottadb.subscript_next("^myglobal", ("sub1", ""))
print(subscript)  # Prints 'sub2'
while True:
    try:
        print(yottadb.subscript_next("^myglobal", ("sub1", subscript)))  # Prints 'sub3', 'sub4', and 'sub5', successively
    except yottadb.YDBNodeEnd:
        break

# subscript_next() also works with subscripts that include data that is not ASCII or valid UTF-8
yottadb.set("mylocal", (b"sub1\x80",)), "val1"),  # Test subscripts with byte strings that are not ASCII or valid UTF-8
yottadb.set("mylocal", (b"sub2\x80", "sub7")), "val2"),
yottadb.set("mylocal", (b"sub3\x80", "sub7")), "val3"),
yottadb.set("mylocal", (b"sub4\x80", "sub7")), "val4"),
print(yottadb.subscript_next(varname="mylocal", subsarray=("",)))  # Prints b"sub1\x80"
print(yottadb.subscript_next(varname="mylocal", subsarray=("sub1\x80",)))  # Prints b"sub2\x80"
print(yottadb.subscript_next(varname="mylocal", subsarray=("sub2\x80",)))  # Prints b"sub3\x80"
print(yottadb.subscript_next(varname="mylocal", subsarray=("sub3\x80",)))  # Prints b"sub4\x80"
try:
    print(yottadb.subscript_next(varname="mylocal", subsarray=("sub4\x80",)))
except YDBNodeEnd:
    pass

Python subscript_previous()

def subscript_previous(varname: AnyStr, subsarray: Tuple[AnyStr] = ()) -> bytes

As a wrapper for the C function ydb_subscript_previous_s() / ydb_subscript_previous_st(), subscript_previous() facilitates reverse breadth-first traversal of a local or global variable sub-tree. A node or subtree does not have to exist at the specified key.

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the subscript level is zero, and variable names should be iterated over instead of subscripts.

  • If there is a previous subscript with a node and/or a subtree, it returns the subscript at the level of the last subscript in subsarray

  • If there is no next node or subtree at that level of the subtree, a yottadb.YDBNodeEnd exception will be raised.

  • If the underlying ydb_subscript_previous_s() / ydb_subscript_previous_st() call returns any other error, the function raises an exception containing the error code and message.

In the special case where subsarray is empty subscript_previous() returns the name of the previous global or local variable, and raises a yottadb.YDBNodeEnd exception if there is no global or local variable preceding varname.

yottadb.set("^myglobal", ("sub1", "sub2"), "val1")
yottadb.set("^myglobal", ("sub1", "sub3"), "val2")
yottadb.set("^myglobal", ("sub1", "sub4"), "val3")
yottadb.set("^myglobal", ("sub1", "sub5"), "val4")

# Get last subscript of the second subscript level
subscript = yottadb.subscript_previous("^myglobal", ("sub1", ""))
print(subscript)  # Prints 'sub5'
while True:
    try:
        print(yottadb.subscript_previous("^myglobal", ("sub1", subscript)))  # Prints 'sub4', 'sub3', and 'sub2', successively
    except yottadb.YDBNodeEnd as e:
        break

Python subscripts()

def subscripts(varname: AnyStr, subsarray: Tuple[AnyStr] = ()) -> SubscriptsIter

The subscripts() function provides a convenient, Pythonic interface for iteratively performing breadth-first traversals at the specified subscript level, starting from the given YottaDB local or global variable node, as specified by the varname and subscripts arguments.

Specifically, subscripts() returns a Python SubscriptsIter iterator object that yields a bytes object representing the next subscript at the given subscript level on each iteration, in accordance with the behavior for Python subscript_next().

Similarly, the reversed version of the returned SubscriptsIter iterator will yield a bytes object representing the previous subscript at the given subscript level on each iteration, in accordance with the behavior for Python subscript_previous().

  • If subsarray is omitted, an empty Tuple is passed by default, signifying that the variable name node should be referenced without any subscripts. In this case, subscripts() will iterate over every local or global variable in the database starting from the local or global variable name specified.

  • If there is a next subscript for a given iteration, the SubscriptsIter iterator will return the subscript at that subscript level as a Python bytes object.

  • If this iterator is passed to the next() built-in function and there is no subscript following the subscript previously returned, a StopIteration exception will be raised.

  • If the underlying Python subscript_next() or Python subscript_previous() call returns any other error, the SubscriptsIter iterator will raise an exception containing the error code and message.

subs = [b"sub1", b"sub2", b"sub3"]
# Set various nodes in the database
for sub in subs:
    yottadb.set("^myglobal", (sub,), str(subs.index(sub)))

# Iterate over all subscripts under a global variable
for subscript in yottadb.subscripts("^myglobal", ("",)):
    print(subscript)  # Prints b'sub1', b'sub2', b'sub3'

# Iterate over some subscripts under a global variable
for subscript in yottadb.subscripts("^myglobal", ("sub1",)):
    print(subscript)  # Prints b'sub2', b'sub3'

# Iterate over all subscripts under a global variable, in reverse
for subscript in reversed(yottadb.subscripts("^myglobal", ("",))):
    print(subscript)  # Prints b'sub3', b'sub2', b'sub1'

# Iterate over some subscripts under a global variable, in reverse
for subscript in reversed(yottadb.subscripts("^myglobal", ("sub3",))):
    print(subscript)  # Prints b'sub2', b'sub1'

Python switch_ci_table()

def switch_ci_table(handle: int) -> int

As a wrapper for the C function ydb_ci_tab_open(), the switch_ci_table() function enables switching of call-in tables by allowing users to switch to a call-in table previously opened by open_ci_table(), as specified through an integer handle argument. This argument should be the return value of a previous call to open_ci_table().

switch_ci_table() returns an integer handle to the previously active call-in table, None if there was none. Switching the call-in table does not change $zroutines, so application code will need to change $zroutines appropriately if the new call-in table requires a different M routine search path.

If the underlying ydb_ci_tab_open() call returns an error, the function raises an exception containing the error code and message.

For an example of how to use switch_ci_table(), see the entry for Python ci() or Python cip().

Python tp()

def tp(callback: object, args: tuple = None, transid: str = "", varnames: Tuple[AnyStr] = None, **kwargs,)

As a wrapper for the C function ydb_tp_s() / ydb_tp_st(), tp() provides an interface for performing basic YottaDB transaction processing from Python code. Specifically, tp() allows users of the Python wrapper to safely call user-defined Python functions containing transaction logic that modifies or updates one or more nodes within a YottaDB database.

A function implementing logic for a transaction should raise one of the following YDBPython exceptions depending on the scenario encountered during transaction processing:

  • If args is not specified, None is passed by default.

  • If transid is not specified, the empty string is passed by default.

  • If varnames is not specified, None is passed by default.

  • When application logic successfully completes execution, no exception should be raised and the transaction can be committed. The YottaDB database engine will commit the transaction if it is able to and, if not, it will call the function again.

  • YDBTPRestart is raised to indicate that the transaction should restart, either because application logic has so determined or because a YottaDB function called by the function has returned YDB_TP_RESTART.

  • YDBTPRollback is raised to indicate that tp() should not commit the transaction, and should raise a YDBTPRollback to the caller.

  • If the underlying ydb_tp_s() / ydb_tp_st() call returns any other error, the function raises an exception containing the error code and message.

The varnames list passed to the tp() method is a list of local variables whose values should be saved, and restored to their original values when the transaction restarts. If the varnames is None, no local variables are saved and restored. If varnames contains one element and that sole element is the string “*” all local variables are saved and restored.

A case-insensitive value of “BA” or “BATCH” for transid indicates to YottaDB that it need not ensure Durability for this transaction (it continues to ensure Atomicity, Consistency, and Isolation), as discussed under ydb_tp_s() / ydb_tp_st().

Please see both the description of ydb_tp_s() / ydb_tp_st() and the section on Transaction Processing for details.

Note

If the transaction logic receives a YDB_TP_RESTART or YDB_TP_ROLLBACK from a YottaDB function or method that it calls, it must return that value to the calling tp() function. Failure to do so could result in application level data inconsistencies and hard to debug application code.

The following example demonstrates a simple usage of tp(). Specifically, a simple callback() function is defined, then wrapped in a simple wrapper() function that calls callback() using tp(), ensuring database integrity via transaction processing. Then, several processes executing the wrapper() function are spawned, each of which attempts to increment the same global variable nodes at once. Each of these processes continues trying to increment the nodes until the incrementation is successful, i.e. YDBTPRestart is not raised. Finally, these processes are gracefully terminated and the values of the global variable nodes are checked to ensure to success of the incrementation attempts of each wrapper() process.

# Define a simple callback function that attempts to increment the global variable nodes represented
# by the given Key objects. If a YDBTPRestart is encountered, the function will retry the continue
# attempting the increment operation until it succeeds.
def callback(fruit1: yottadb.Key, fruit2: yottadb.Key, fruit3: yottadb.Key) -> int:
    while True:
        try:
            fruit1.incr()
            fruit2.incr()
            fruit3.incr()
            break
        except yottadb.YDBTPRestart:
            continue

    return yottadb.YDB_OK

# Define a simple wrapper function to call the callback function via tp().
# This wrapper will then be used to spawn multiple processes, each of which
# calls tp() using the callback function.
def wrapper(function: Callable[..., object], args: Tuple[AnyStr]) -> int:
    return yottadb.tp(function, args=args)

# Create keys
apples = yottadb.Key("^fruits")["apples"]
bananas = yottadb.Key("^fruits")["bananas"]
oranges = yottadb.Key("^fruits")["oranges"]
# Initialize nodes
apples_init = "0"
bananas_init = "5"
oranges_init = "10"
apples.value = apples_init
bananas.value = bananas_init
oranges.value = oranges_init

# Spawn some processes that will each call the callback function
# and attempt to access the same nodes simultaneously. This will
# trigger YDBTPRestarts, until each callback function successfully
# updates the nodes.
num_procs = 10
processes = []
for proc in range(0, num_procs):
    # Call the callback function that will attempt to update the given nodes
    process = multiprocessing.Process(target=wrapper, args=(callback, (apples, bananas, oranges)))
    process.start()
    processes.append(process)
# Gracefully terminate each process and confirm it exited without an error
for process in processes:
    process.join()
    assert process.exitcode == 0

# Confirm all nodes incremented by num_procs, i.e. by one per callback process spawned
assert int(apples.value) == int(apples_init) + num_procs
assert int(bananas.value) == int(apples_init) + num_procs
assert int(oranges.value) == int(apples_init) + num_procs

Python transaction()

def transaction(function) -> Callable[..., object]

The transaction() function is provided as a decorator for convenience to simplify the basic case of passing a callback function to Python tp() when no special handling is needed. It is not intended to be used on its own, but instead for decorating functions that require transaction processing. Users with more sophisticated transaction processing needs are encouraged to write their own decorator functions for handling transactions.

transaction() converts the specified function into a form safe for use in YottaDB database transactions. Specifically, it wraps function in a new function definition that includes a call to Python tp() and basic transaction exception handling. This new wrapper function is then returned and may then be used as a transaction-safe version of the passed function. Accordingly, function should be written as if it were to be passed to Python tp().

Since this function simply wraps the passed function in a new function definition, it will always succeed. However, the resulting wrapper function may raise exceptions depending on its execution. For more information about this behavior, see the entry for Python tp(), as the wrapper function is a pre-populated call to this function.

  • If the wrapped function returns None, then yottadb.YDB_OK will be returned to the wrapping Python tp() call

  • If the wrapped function returns any other value, this value will be returned directly to the wrapping Python tp() call without modification

  • If the wrapped function raises yottadb.YDBTPRestart, then yottadb.YDB_TP_RESTART will be returned to the wrapping Python tp() call

# Wrap a simple function with the transaction
@yottadb.transaction
def my_transaction(key1: yottadb.Key, value1: str, key2: yottadb.Key, value2: str) -> None:
     key1.value = value1
     key2.value = value2

# Create Key objects to pass to the newly defined and decorated my_transaction() function
key1 = yottadb.Key("^myglobal")["sub1"]["sub2"]
key2 = yottadb.Key("^myglobal")["sub1"]["sub3"]

# Call the function decorated with transaction()
status = my_transaction(key1, "val1", key2, "val2")
# Handle possible results of the call as one would handle results of a call to tp()
if yottadb.YDB_OK == status:
    # Transaction successful
    print(key1.value)  # Prints 'val1'
    print(key2.value)  # Prints 'val2'
else if yottadb.YDB_TP_RESTART == status:
    # Restart the transaction
    print(status)
else if yottadb.YDB_TP_ROLLBACK == status:
    # Do not commit the transaction
    print(status)
else:
    # Another error occurred
    # Do not commit the transaction
    print(status)

Python zwr2str()

def zwr2str(string: AnyStr) -> bytes

As a wrapper for the C function ydb_zwr2str_s() / ydb_zwr2str_st(), zwr2str takes a string in ZWRITE format and returns it as a regular string. This method is the inverse of Python str2zwr().

Note that the return value of this function is always a bytes object, reflecting the fact that YottaDB stores all values as binary data, such that a global or local variable node value is not guaranteed to be a valid UTF-8 string. Accordingly, the return value of this function is not guaranteed to be castable to a Python str object.

print(yottadb.zwr2str(b'"X"_$C(0)_"ABC"'))  # Prints b'X\x00ABC'

YottaDB Key class properties

Key.data

@property
def data(self) -> int

Matching Python data(), the Key.data property method returns the result of ydb_data_s() / ydb_data_st() (0, 1, 10, or 11).

In the event of an error in ydb_data_s() / ydb_data_st(), a YDBError exception is raised reflecting YottaDB error code and message.

key = yottadb.Key("mylocal")["sub1"]["sub2"]
key.value = "test"
print(key.data) # Prints 1
print(key.parent.data) # Prints 10
print(key["sub3"].data) # Prints 0
key["sub3"].value = "test2"
print(key["sub3"].data) # Prints 1
print(key.data) # Prints 11

Key.has_value

@property
def has_value(self) -> bool

Key.has_value provides a class property that returns True or False depending on whether the global or local variable node represented by the given Key object has a value or does not have a value, respectively.

In the event of an error in the underlying ydb_data_s() / ydb_data_st() call, a YDBError exception is raised reflecting the error code and message.

This property references Key.data internally, and is provided for convenience.

key = yottadb.Key("mylocal")["sub1"]["sub2"]
print(key.has_value) # Prints False
key.value = "test"
print(key.has_value) # Prints True

Key.has_tree

@property
def has_tree(self) -> bool

Key.has_tree provides a class property that returns True or False depending on whether the global or local variable node represented by the given Key object has a (sub)tree or does not have a (sub)tree, respectively.

In the event of an error in the underlying ydb_data_s() / ydb_data_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

This property references Key.data internally, and is provided for convenience.

key = yottadb.Key("mylocal")["sub1"]["sub2"]
key.value = "test"
print(key.has_tree) # Prints False
print(key.parent.has_tree) # Prints True

Key.subsarray

@property
def subsarray(self) -> List[AnyStr]

Key.subsarray provides a class property that returns the subscripts of the global or local variable node represented by the given Key object as a List of str or bytes objects, depending on whether the Key was constructed using str or bytes objects to specify the variable name or subscripts.

key = yottadb.Key("mylocal")["sub1"]["sub2"]
print(key.subsarray) # Prints ["sub1", "sub2"]

Key.subscripts

@property
def subscripts(self) -> Generator

Key.subscripts provides a class property that returns a Generator for iterating over subscripts at the level of the global or local variable node represented by the given Key object. Each iteration will yield the result of a call to subscript_next, i.e. a bytes object representing a YottaDB subscript.

In the event of an error in an underlying ydb_subscript_next_s() / ydb_subscript_next_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

Example

key = yottadb.Key("^myglobal")["sub1"]["sub2"]
for subscript in key.subscripts:
    print(subscript)  # Prints the next subscript at the "sub2" subscript level of the key

Key.subsarray_keys

@property
def subsarray_keys(self) -> List["Key"]:

Key.subsarray_keys provides a class property that returns the subscripts of the global or local variable node represented by the given Key object as a List of other Key objects. Each of these Key objects represents a full YottaDB global or local variable node (variable name and subscripts).

key = yottadb.Key("mylocal")["sub1"]["sub2"]
print(key.subsarray_keys) # Prints [Key:mylocal("sub1"), Key:mylocal("sub1","sub2")]

Key.value

@property
def value(self) -> Optional[AnyStr]

@value.setter
def value(self, value: AnyStr) -> None

Acting as a class property, Key.value wraps both ydb_get_s() / ydb_get_st() and ydb_set_s() / ydb_set_st() to set or get the value at the global or local variable node or intrinsic special variable represented by the given Key object.

In the event of an error in the underlying ydb_get_s() / ydb_get_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

Example:

key = yottadb.Key("^myglobal")
key.value = "such wow"
print(key.value)  # Prints "such wow"

Key.varname_key

@property
def varname_key(self) -> Optional["Key"]:

Key.varname_key provides a class property that returns a Key object for the unsubscripted global or local variable node represented by the given Key object as a str object.

key = yottadb.Key("mylocal")["sub1"]["sub2"]
print(key.varname_key) # Prints Key:mylocal

Key.varname

@property
def varname(self) -> AnyStr

Key.varname provides a class property that returns the name of the global or local variable node represented by the given Key object as a bytes or str object, depending on how the Key variable name was specified.

key = yottadb.Key("mylocal")["sub1"]["sub2"]
print(key.varname) # Prints 'mylocal'

YottaDB Key class regular methods

Key.delete_node()

def delete_node(self) -> None

Matching Python delete_node(), Key.delete_node() wraps ydb_delete_s() / ydb_delete_st() with a value of YDB_DEL_NODE for deltype to delete a local or global variable node, specifying that only the node should be deleted, leaving the (sub)tree untouched.

In the event of an error in the underlying ydb_delete_s() / ydb_delete_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

key = yottadb.Key("mylocal")["sub1"]["sub2"]
key.value = "test"
print(key.value) # Prints b'test'
key.delete_node()
print(key.value) # Prints None

Key.delete_tree()

def delete_tree(self) -> None

Matching Python delete_tree(), Key.delete_tree() wraps ydb_delete_s() / ydb_delete_st() with a value of YDB_DEL_TREE for deltype to delete the local or global variable node represented by the Key object, along with its (sub)tree.

In the event of an error in the underlying ydb_delete_s() / ydb_delete_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

key = yottadb.Key("mylocal")["sub1"]["sub2"]
print(key.data) # Prints 0
key.value = "test"
print(key.data) # Prints 1
print(key.parent.data) # Prints 10
key.parent.delete_tree()
print(key.data) # Prints 0
print(key.parent.data) # Prints 0

Key.get()

def get(self) -> Optional[bytes]

Matching Python get(), Key.get() wraps ydb_get_s() / ydb_get_st() to retrieve the value of the local or global variable node represented by the given Key object, returning it as a bytes object.

In the event of an error in the underlying ydb_get_s() / ydb_get_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

key = yottadb.Key("mylocal")["sub1"]["sub2"]
print(key.get()) # Prints None
key.set("test")
print(key.get()) # Prints b'test'

Key.incr()

def incr(self, increment: Union[int, float, str, bytes] = "1") -> bytes

Matching Python incr(), Key.incr() wraps ydb_incr_s() / ydb_incr_st() to atomically increment the global or local variable node represented by the Key object coerced to a number, with increment coerced to a number. If successful, the call returns the resulting value as a bytes object.

  • If increment is omitted, a value of 1 is used by default.

  • If ydb_incr_s() / ydb_incr_st() returns an error such as NUMOFLOW, an exception will be raised.

  • Otherwise, it increments the specified node and returns the resulting value.

  • In the event of any other error in the underlying ydb_incr_s() / ydb_incr_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

If unspecified, the default increment is 1. Note that the value of the empty string coerced to an integer is zero, but 1 is a more useful default value for an omitted parameter in this case.

key = yottadb.Key("mylocal")["sub1"]["sub2"]
print(key.value) # Prints None
print(key.incr()) # Prints b'1'
print(key.incr()) # Prints b'2'

Key.lock()

def lock(self, timeout_nsec: int = 0) -> None

Matching Python lock(), Key.lock() releases all lock resources currently held and then attempts to acquire the named lock resource represented by the given Key object. In other words, Key.lock() will attempt to acquire a lock for the single key represented by the given Key object.

  • If timeout_nsec is omitted, a value of 0 is used by default.

  • If timeout_nsec exceeds yottadb.YDB_MAX_TIME_NSEC, a yottadb.YDBError exception will be raised where yottadb.YDB_ERR_TIME2LONG == YDBError.code()

  • If it is able to acquire the lock resource within timeout_nsec nanoseconds, it returns holding the lock, otherwise it raises a YDBTimeoutError exception. If timeout_nsec is zero, the function makes exactly one attempt to acquire the lock.

  • If the underlying ydb_lock_s() / ydb_lock_st() call returns any other error, the function raises a YDBError exception containing the error code and message.

The following example provides a demonstration of basic Key locking operations. The example locks the given Key, then attempts to increment the lock on it by calling a separately defined lock_value() helper function as a separate Python process. Due to the initial locking of the key, this lock_value() fails with an exit code of 1. Next, all locks are released and a new lock_value() process is spawned that again attempts to increment the lock on the key. Since all locks were previously released, this new attempt succeeds and the process exits with a 0 exit code.

import multiprocessing
import datetime

# Lock a value in the database
def lock_value(key: Union[yottadb.Key, tuple], interval: int = 2, timeout: int = 1):
    if isinstance(key, yottadb.Key):
        varname = key.varname
        subsarray = key.subsarray
    else:
        varname = key[0]
        subsarray = key[1]
    if len(subsarray) == 0:
        subsarray = None

    has_lock = False
    try:
        yottadb.lock_incr(varname, subsarray, timeout_nsec=(timeout * 1_000_000_000))
        print("Lock Success")
        has_lock = True
    except yottadb.YDBTimeoutError:
        print("Lock Failed")
        sys.exit(1)
    except Exception as e:
        print(f"Lock Error: {repr(e)}")
        sys.exit(2)

    if has_lock:
        time.sleep(interval)
        yottadb.lock_decr(varname, subsarray)
        if timeout != 0 or interval != 0:
            print("Lock Released")

    sys.exit(0)


key = yottadb.Key("^test4")["sub1"]["sub2"]
# Attempt to get the lock
key.lock()
# Attempt to increment/decrement the lock
process = multiprocessing.Process(target=lock_value, args=(key,))
process.start()
process.join()
print(process.exitcode)  # Prints 1
# Release all locks
yottadb.lock()
# Attempt to increment/decrement the lock
process = multiprocessing.Process(target=lock_value, args=(key,))
process.start()
process.join()
print(process.exitcode)  # Prints 0

Key.lock_decr()

def lock_decr(self) -> None

Matching Python lock_decr() Key.lock_decr() wraps ydb_lock_decr_s() / ydb_lock_decr_st() to decrement the count of the lock name represented by the given Key object, releasing it if the count goes to zero or ignoring the invocation if the process does not hold the lock.

In the event of an error in the underlying ydb_lock_decr_s() / ydb_lock_decr_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

import multiprocessing
import datetime

key = yottadb.Key("^myglobal")["sub1"]
# For the definition of lock_value(), see the entry for Key.lock()
process = multiprocessing.Process(target=lock_value, args=(key,))
process.start()
time.sleep(0.5)  # Wait for new process to spawn

t1 = datetime.datetime.now()
yottadb.Key("mylocal").lock_incr()
t2 = datetime.datetime.now()

time_elapse = t2.timestamp() - t1.timestamp()
print(time_elapse)  # Prints number of seconds elapsed
key.lock_decr()
time.sleep(0.5)  # Wait for lock to release
process.join()

Key.lock_incr()

def lock_incr(self, timeout_nsec: int = 0) -> None

Matching Python lock_incr(), Key.lock_incr() wraps ydb_lock_incr_s() / ydb_lock_incr_st() to attempt to acquire the lock resource name represented by the given Key object without releasing any locks the process already holds.

  • If timeout_nsec is omitted, a value of 0 is used by default.

  • If the process already holds the named lock resource, the method increments its count and returns.

  • If timeout_nsec exceeds yottadb.YDB_MAX_TIME_NSEC, the method raises a TIME2LONGError exception.

  • If it is able to acquire the lock resource within timeout_nsec nanoseconds, it returns holding the lock, otherwise it raises a YDBTimeoutError exception. If timeout_nsec is zero, the method makes exactly one attempt to acquire the lock.

For an example of how to use this function, see Key.lock_decr().

Key.node_next()

def node_next(varname: AnyStr, subsarray: Tuple[AnyStr] = ()) -> Tuple[bytes, ...]

Matching Python node_next(), Key.node_next() wraps ydb_node_next_s() / ydb_node_next_st() to facilitate depth first traversal of the local or global variable tree represented by the given Key object.

  • If there is a next node, it returns the subscripts of that next node as a tuple of Python bytes objects.

  • If there is no node following the specified node, a yottadb.YDBNodeEnd exception will be raised.

  • In the event of an error in the underlying ydb_node_next_s() / ydb_node_next_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

Key.set()

def set(self, value: AnyStr = "") -> None

Matching Python set(), Key.set() wraps ydb_set_s() / ydb_set_st() to set the local or global variable node represented by the given Key object to the value specified by value.

In the event of an error in the underlying ydb_set_s() / ydb_set_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

key = yottadb.Key("mylocal")["sub1"]["sub2"]
print(key.get()) # Prints None
key.set("test")
print(key.get()) # Prints b'test'

Key.subscript_next()

def subscript_next(self, reset: bool = False) -> bytes

Matching Python subscript_next(), Key.subscript_next() wraps ydb_subscript_next_s() / ydb_subscript_next_st() to facilitate breadth-first traversal of the local or global variable sub-tree at the subscript level represented by the given Key object. A node or subtree does not have to exist at the specified key. The reset parameter may be used to instruct Key.subscript_next() to begin traversal at the first subscript at the current subscript level, even if Key.subscript_next() has already traversed over it.

  • If reset is omitted, it is set to False by default.

  • At the level of the last subscript, if there is a next subscript with a node and/or a subtree that subscript will be returned as a bytes object.

  • If there is no next node or subtree at that level of the subtree, a yottadb.YDBNodeEnd exception will be raised.

  • A yottadb.YDBNodeEnd exception will be raised on all subsequent calls to Key.subscript_next() after exhausting all nodes and/or subtrees as described above

  • To enable re-traversal of the current subscript level, the user may pass a value of True to Key.subscript_next(), which will cause the function to return the next subscript at the current level, as if Key.subscript_next() was not previously called and a yottadb.YDBNodeEnd exception was not previously raised.

  • In the event of any other error in the underlying ydb_subscript_next_s() / ydb_subscript_next_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

The following example sets a value on multiple nodes at the first subscript level of a local variable, then iterates over each subscript at this level in two ways. First, the subscripts are iterated over using a Key.subscript_next() manually in a succession of hard-coded calls. Then, the starting subscript of the iteration is reset after iterating over all subscripts at that level. Finally, the subscripts are again iterated over, but this time using a while loop instead of hard-coded individual calls to Key.subscript_next().

key = yottadb.Key("testsubsnext")
key["sub1"] = "1"
key["sub2"] = "2"
key["sub3"] = "3"
key["sub4"] = "4"

print(key.subscript_next())  # Prints "sub1"
print(key.subscript_next())  # Prints "sub2"
print(key.subscript_next())  # Prints "sub3"
print(key.subscript_next())  # Prints "sub4"

try:
    key.subscript_next()
except yottadb.YDBNodeEnd:
    print(key[key.subscript_next(reset=True)].value)  # Prints b"1"
    print(key[key.subscript_next()].value)  # Prints b"2"
    print(key[key.subscript_next()].value)  # Prints b"3"
    print(key[key.subscript_next()].value)  # Prints b"4"

try:
    sub = key.subscript_next(reset=True)  # Resets starting subscript to ""
except yottadb.YDBNodeEnd:
    # There are subscripts defined for the given Key, so a reset of subscript_next's
    # next subscript to the default starting subscript of "" should not return
    # a YDBError of YDB_ERR_NODEEND. If, on the other hand, there were no subscripts for the
    # given Key, subscript.next() would always raise a YDBError of YDB_ERR_NODEEND, regardless of
    # whether the `reset` argument is set to True or not.
    assert False

count = 1
print(sub)  # Prints "sub1"
while True:
    try:
        sub = key.subscript_next()
        count += 1
        assert sub == "sub" + str(count)
    except yottadb.YDBNodeEnd:
        break

Key.subscript_previous()

def subscript_previous(self, reset: bool = False) -> bytes

Matching Python subscript_previous(), Key.subscript_previous() wraps ydb_subscript_previous_s() / ydb_subscript_previous_st() to facilitate reverse breadth-first traversal of the local or global variable sub-tree at the subscript level represented by the given Key object. A node or subtree does not have to exist at the specified key.

  • If reset is omitted, it is set to False by default.

  • At the level of the last subscript, if there is a previous subscript with a node and/or a subtree that subscript will be returned as a bytes object.

  • If there is no previous node or subtree at that level of the subtree, a yottadb.YDBNodeEnd exception will be raised.

  • In the event of an error in the underlying ydb_subscript_previous_s() / ydb_subscript_previous_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

The following example sets a value on multiple nodes at the first subscript level of a local variable, then iterates over each subscript at this level in two ways. First, the subscripts are iterated over using a Key.subscript_previous() manually in a succession of hard-coded calls. Then, the starting subscript of the iteration is reset after iterating over all subscripts at that level. Finally, the subscripts are again iterated over, but this time using a while loop instead of hard-coded individual calls to Key.subscript_previous().

key = yottadb.Key("testsubsprevious")
key["sub1"] = "1"
key["sub2"] = "2"
key["sub3"] = "3"
key["sub4"] = "4"

print(key.subscript_previous())  # Prints "sub4"
print(key.subscript_previous())  # Prints "sub3"
print(key.subscript_previous())  # Prints "sub2"
print(key.subscript_previous())  # Prints "sub1"

try:
    key.subscript_previous()
except yottadb.YDBNodeEnd:
    print(key[key.subscript_previous(reset=True)].value)  # Prints b"4"
    print(key[key.subscript_previous()].value)  # Prints b"3"
    print(key[key.subscript_previous()].value)  # Prints b"2"
    print(key[key.subscript_previous()].value)  # Prints b"1"

try:
    sub = key.subscript_previous(reset=True)  # Resets starting subscript to ""
except yottadb.YDBNodeEnd:
    # There are subscripts defined for the given Key, so a reset of subscript_previous's
    # previous subscript to the default starting subscript of "" should not return
    # a YDBError of YDB_ERR_NODEEND. If, on the other hand, there were no subscripts for the
    # given Key, subscript.previous() would always raise a YDBError of YDB_ERR_NODEEND, regardless of
    # whether the `reset` argument is set to True or not.
    assert False

count = 4
print(sub)  # Prints "sub4"
while True:
    try:
        sub = key.subscript_previous()
        count -= 1
        assert sub == "sub" + str(count)
    except yottadb.YDBNodeEnd as e:
        break

YottaDB Key class magic methods

Key.__eq__()

def __eq__(self, other) -> bool

The Key.__eq__() magic method allows for easy comparison between two Key objects, using the Python == operator. If the two Key objects represent the same YottaDB local or global variable node, then Key.__eq__() will return True, otherwise it will return False. For example:

key = yottadb.Key("^myglobal")["sub1"]["sub2"]
key2 = yottadb.Key("^myglobal")["sub1"]["sub2"]
print(key == key2) # Prints True

key = yottadb.Key("^myglobal")["sub1"]["sub2"]
key2 = yottadb.Key("^myglobal")["sub1"]
print(key == key2) # Prints False

Key.__getitem__()

def __getitem__(self, item)

The Key.__getitem__() magic method creates a new Key object by adding the specified item as an additional subscript on the given Key object.

This enables usage of the standard index bracket syntax ([]) for the transparent production of new Key objects for both in-line, one-off usage and for the creation of new objects for later use.

For example:

key1 = yottadb.Key("^myglobal")
key2 = key1["sub1"]
key3 = key2["sub2"]
key4 = key2["sub3"]
print(str(key1)) # Prints '^myglobal'
print(str(key2)) # Prints '^myglobal("sub1")'
print(str(key3)) # Prints '^myglobal("sub1","sub2")'
print(str(key4)) # Prints '^myglobal("sub1","sub3")'

Key.__iadd__()

def __iadd__(self, num: Union[int, float, str, bytes]) -> Optional["Key"]

The Key.__iadd__() magic method allows for easy incrementation of the YottaDB local or global variable node represented by the Key object, using the Python += operator. For example:

In the event of an error in the underlying ydb_incr_s() / ydb_incr_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

key = yottadb.Key("^myglobal")["sub1"]["sub2"]
key.value = 2
key += 2
print(key.value) # Prints '4'

Key.__init__()

def __init__(self, name: AnyStr, parent: Key = None) -> None

The Key.__init__() function acts as the constructor for the Key class and is used to create new Key objects.

Note: Users should not attempt to set parent, but omit this parameter. This is because parent is used implicitly by several Key methods and not intended for use by users. If a user nonetheless passes a valid parent argument, i.e. a Key object, then a new Key will be generated where name is appended as an additional subscript at the end of the subscript array of the parent Key.

The following errors are possible during Key creation: - TypeError: when name is not of type bytes or str. - TypeError: when parent is not of type Key or None. - ValueError: if a subscript array is specified for a YottaDB Intrinsic Special Variable (ISV), i.e. parent is not None and name specifies an ISV. - ValueError: if the subscript array, passed via parent, exceeds yottadb.YDB_MAX_SUBS in length.

key = yottadb.Key("^myglobal")["sub1"]["sub2"]

# Set `parent` explicitly (not recommended)
key = yottadb.Key("sub3", parent=key)  # Raises TypeError for non-Key `parent` argument
print(str(key)) # Prints '^myglobal("sub1","sub2","sub3")'
key = yottadb.Key("^myglobal", parent="not a key object")  # Raises TypeError for non-Key `parent` argument

# Proper ISV Key creation
key = yottadb.Key("$ZSTATUS")
print(str(key)) # Prints '$ZSTATUS'
# Invalid ISV Key creation
key = yottadb.Key("$ZSTATUS")["sub1"]["sub2"]  # Raises ValueError for subscripted ISV

Key.__isub__()

def __isub__(self, num: Union[int, float, str, bytes]) -> Optional["Key"]

The Key.__isub__() magic method allows for easy decrementation of the YottaDB local or global variable node represented by the Key object, using the Python -= operator. For example:

In the event of an error in the underlying ydb_incr_s() / ydb_incr_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

key = yottadb.Key("^myglobal")["sub1"]["sub2"]
key.value = 2
key -= 2
print(key.value) # Prints '0'

Key.__iter__()

def __iter__(self) -> Generator

The Key.__iter__() magic method allows for easy iteration over the subscripts at the subscript level of the given Key object, beginning from the first subscript. Accordingly, this method is useful for breadth-first searches. For example,

In the event of an error in an underlying ydb_subscript_next_s() / ydb_subscript_next_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

fruits = yottadb.Key("^inventory")["fruits"]
fruits["apples"] = 'in stock'
fruits["bananas"] = 'in stock'
fruits["oranges"] = 'sold out'
fruits["kiwis"] = 'in stock'
fruits["398576986"] = "in stock"
fruits["839587329"] = "sold out"
fruits[b"\x80pomegranates"] = b"\x80sold out"  # byte strings that are not ASCII or valid UTF-8 are also supported
for fruit in fruits:
    print(f"{fruit}: {fruit.value}")

# Prints:
# inventory("fruits",398576986): b'in stock'
# inventory("fruits",839587329): b'sold out'
# inventory("fruits","apples"): b'in stock'
# inventory("fruits","bananas"): b'in stock'
# inventory("fruits","kiwis"): b'in stock'
# inventory("fruits","oranges"): b'sold out'
# inventory("fruits",$ZCH(128)_"pomegranates"): b'\x80sold out'

Key.__repr__()

def __repr__(self) -> str

The Key.__repr__() magic method returns a Python-readable representation of the Key object. Specifically, :code`Key.__repr__()` produces a representation of the Key object that can be passed to the built-in eval() function to produce a new instance of the object.

Note, however, that this cannot be done with perfect reliability, as successful object reproduction will depend on how the yottadb module is imported. To provide flexibility, Key.__repr__() produces a representation as if the Key class is imported directly, i.e. from yottadb import Key. This allows for eval() to be used to reproduce a Key object, provided that the str passed to it includes any module import prefixes qualifying the Key name. For example:

import yottadb

key = yottadb.Key("^myglobal")["sub1"]["sub2"]
print(repr(key)) # Prints Key("^myglobal")["sub1"]["sub2"]

# Attempt to call eval() without fully qualifying the import prefix for the Key class
try:
    eval(repr(key))
except NameError:
    # eval() raises: "NameError: name 'Key' is not defined"
    assert True

# Call eval() with a fully qualified import prefix for the Key class
print(repr(eval("yottadb." + repr(key))))  # Prints Key("^myglobal")["sub1"]["sub2"]

Key.__reversed__()

def __reversed__(self) -> Generator

The Key.__reversed__() magic method allows for easy iteration over the subscripts at the subscript level of the given Key object, beginning from the last subscript. Accordingly, this method is useful for breadth-first searches, in reverse order. For example,

In the event of an error in an underlying ydb_subscript_previous_s() / ydb_subscript_previous_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

vegetables = yottadb.Key("^inventory")["vegetables"]
vegetables["carrots"] = 'in stock'
vegetables["cabbages"] = 'sold out'
vegetables["potatoes"] = 'in stock'
vegetables["spinach"] = 'sold out'
for vegetable in reversed(vegetables):
    print(f"{vegetable}: {vegetable.value}\n")

# Prints:
# spinach: sold out
# potatoes: in stock
# cabbages: sold out
# carrots: in stock

Key.__setitem__()

def __setitem__(self, item, value)

The Key.__setitem__() magic method provides a simple interface for updating the value at the YottaDB local or global variable node represented by the Key object, using the Python = operator. For example:

In the event of an error in the underlying ydb_set_s() / ydb_set_st() call, a YDBError exception is raised reflecting the underlying YottaDB error code and message.

key = yottadb.Key("^myglobal")["sub1"]
key["sub2"] = "my value"
print(key["sub2"].value) # Prints 'my value'

Key.__str__()

def __str__(self) -> str

The Key.__str__() magic method returns a human-readable representation of the Key object as a Python str object. For a Python-readable representation of the object, use Key.__repr__().

key = yottadb.Key("^myglobal")["sub1"]["sub2"]
print(str(key)) # Prints '^myglobal("sub1","sub2")'