Project Description

This package has collection of functions which can be used to compare two dataframes. These functions can help in validation, where there is a need to compare if any of the metrics value has been altered or to get matching records in the two dataframes. The package serves best when the two dataframes have same schema and from the same source. In validation, if you are comparing two datas which may have changed over the time, this package can help you in the analysis and save the time spent in BI Tools for the analysis.

IMPORT:

from compare_df import *

Function I : getMatchingRecords(dataframe1=None,dataframe2=None,common_columns=None)

This function returns a dataframe which contains identical records for the passed list of columns (default :- all columns) in the two dataframes.

The Parameters:-

:param dataframe1: The first Input DataFrame(X) :param dataframe2:The second Input DataFrame(X) :param common_columns: The list of columns for which the two dataframes have to be compared. default : All columns of dataframe1

Return

This Returns a dataframe which holds the records which have identical value for the passed columns in both dataframes.

Example:

from compare_df import * getMatchingRecords(df1,df2,common_columns=["Prduct_no","Product_Category"])

This returns df with the records of Values of columns ["Prduct_no","Product_Category"] which were found in both dataframes.

Function 2 : getUniqueRecords(dataframe1=None,dataframe2=None,key=[])

This function receives two dataframes and list of key attributes. These attributes should identify a record in both dataframe uniquely. It returns a dataframe with those records which are only present in either of the dataframe. The dataframe has a column named 'Dataframe' which tells in which dataframe does the record exist

The Parameters:-

:param dataframe1: The first Input DataFrame(X), also referred as left. :param dataframe2:The second Input DataFrame(X), also referred as right. :param key: The list of columns present in both dataframes which must identify a record in both dataframes uniquely.

Return

This Returns a dataframe which holds the records which are uniquely present in the dataframe along with the dataframe name.

Example:

from compare_df import * getMatchingRecords(df1,df2,common_columns=["Department","Emp_id"])

This returns df for which the key attributes : ["Prduct_no","Product_Category"] are present in only one of the dataframe.

Function 3 : getVariableRecords(dataframe1, dataframe2, key, metrics=[], unique=True)

This function receives two dataframes and list of key attributes and metric attributes and a boolean attribute -> 'unique'. These attributes should identify a record in both dataframe uniquely. Metric attributes are those attributes for which you want to compare the value between the two dataframes. The function returns the key values and the metric column for which irregularity was encountered. It classify the irregularity into three classes: Mismatch : If key values are present in both dataframe, but there is a mismatch for the column value left_only : If the key values are present only in the first/left dataframe but not in second dataframe. right_only : If the key values are present only in the second/right dataframe but not in first dataframe. If param -> 'unique' is set to True: Only the Records tagged as "Mismatch" are returned If param -> 'unique' is set to False: All the Irregular Records are returned

The Parameters:-

:param dataframe1: Pandas DataFrame : The first Input DataFrame(X), also referred as left. :param dataframe2: Pandas DataFrame : The second Input DataFrame(X), also referred as right. :param key: List :The list of columns present in both dataframes which must identify a record in dataframe uniquely. :param metrics: List :The list of columns for which comparison needs to be made. Default -> Columns in dataframe1 other than the key columns. :param unique : Boolean : If True, returns only the records with Comparison tagged as Mismatch ; If False, returns all the irregularities. Default -> True

Return

This Returns a dataframe which observe a difference in the value of any Metrics column passed between the two dataframes for the respective key values

Example:

from compare_df import * getVariableRecords(df1,df2,key=["Department","Emp_id"],metrics=["ratings","contact"])

The function will check for each column in metrics for all key records, check if both dataframe has same values, If the values are different then the key value, column name, value in the two dataframes would be returned. Along with a Comparison column which state whether there is Mismatch in the values, or the value is present in only one dataframe.