wrappers

A List of Useful Decorators/Wrappers for pandas Modules

The decorators can be used for profiling/understanding the a function which are developed to handle a DataFrame object.

pandaswizard.wrappers.recordCounter(func: callable) → callable

Verbose the Shape of the DataFrame Pre- & Post- Function Execution

On execution of a function (typically, if the function uses the drop() or merge() or etc.) the decorator prints the number of records before and after the execution.

import pandas as pd
import pandaswizard as pdw # attempt to create an ubiquitous naming

frame = pd.DataFrame(data = {
    "LABEL" : ["A", "A", "B"],
    "VALUES" : [1, 2, 3]
})

@pdw.wrappers.recordCounter
def dropvals(frame):
    return frame[frame["LABEL"] != "B"]

filtered = dropvals(frame = frame.copy())
>> Executed with @recordCounter[`dropvals`]
>>   >> Original Record Count = 3
>>   >> Final Record Count    = 2
>>   >> Dropped/Added Records = -1 (= 33.333%)

LIMITATION: The decorator works IFF the dataframe is either the first arguments or is passed as an keyword argument with the argument name as either data or frame which is the general convention that is followed throughout the module. If the code has arguments, then it has preference over the keyword arguments and the dataframe must be the first argument.

USE Cases: Often on function execution, joining/dropping records users need to execute df.shape before and after to check the record count. This is simplified and gives the result, and thus reducing code during such operations.

pandaswizard.wrappers.timeit(func: callable) → callable

A Mimic of the iPython Magic %timeit for Dataframes

The built-in iPython magic function %timeit displays the executed time of a function, or like the one from command line:

python -m timeit "function()"

The function is built specifically to handle functions which returns a pd.DataFrame object, thus it prints more information like the number of records fetched, shape and other information w/o explictly needing to call additional pandas function by the end-user.