Constraints.multiparent_conditions

Function for replacement of values in a dataframe based on multiple conditions evaluated from multiple columns.

Constraints.multiparent_conditions(df, var_array, dict_conditions_values)

Parameters

df: (dataframe)
- dataframe to be updated
var_array: (list)
- A list of strings of column names to be updated.
dict_conditions_values: (dict)
- A dictionary of conditions and values. The conditions are evaluated from multiple columns and the corresponding value is then inserted into the specified columns.

Returns

pandas.DataFrame
- the updated dataframe

Notes

Snippet of an example dict_conditions_values

df = multiparent_conditions(df, ["column1", "column2"], 
    {0: 
        {
            "conditions": {
                "parent1": {"parent": "parent1_column", "condition": "> 5"}, 
                "parent2": {"parent": "parent2_column", "condition": "< 10"}
            }, 
            "value": 0
        }, 
    1: {
            "conditions": {
                "parent3": {"parent": "parent3_column", "condition": "== 'Yes'"}
            }, 
            "value": 1 #note that assigned value must match the dtype of column
        }
    }
)

This example updates the dataframe with the values 0 and 1 in "column1" and "column2" columns, according to the conditions given. In particular, the value 0 is inserted when "parent1_column" is greater than 5 AND "parent2_column" is less than 10. The value 1 is inserted when "parent3_column" is equal to "Yes".

Examples

FIX nPregnancies, nBabies, Age1stBaby (numerical)

# nPregnancies: How many times participant has been pregnant. Reported for female participants aged 20 years or older
# nBabies: How many of participants deliveries resulted in live births. Reported for female participants aged 20 years or older.
# Age1stBaby: Age of participant at time of first live birth. 14 years or under = 14, 45 years or older = 45. Reported for female participants aged 20 years or older.

vArray = ['nPregnancies', 'nBabies', 'Age1stBaby']
age_lessthan20 = {'parent': 'Age', 'condition': '<20'}
genderisMale = {'parent': 'Gender', 'condition': '=="male"'}
dict_conditions_values = {
    1: {
        'conditions': {1: age_lessthan20},
        'value': np.nan
    },
    2: {
        'conditions': {1: genderisMale},
        'value': np.nan
    }
}
df = con.multiparent_conditions(df, vArray, dict_conditions_values)