CODEWITHSUNDEEP
pythonpython-3.xpandasfilterfunctional-programming
KubiK888
KubiK888

Reputation: 4723

How to make flexible command filter for Python Pandas dataframe

I have the following df:

These are the counts of prescribed drug before and after a patient's index date. 

PREINDEX_N_DRUG_A_TAG   POSTINDEX_N_DRUG_A_TAG  PREINDEX_N_DRUG_B_TAG   POSTINDEX_N_DRUG_B_TAG  PREINDEX_N_DRUG_C_TAG   POSTINDEX_N_DRUG_C_TAG
                    0                       8                   0                   8                       0                       8
                    0                       2                   4                   0                       8                       4
                    2                       0                   0                   4                       3                       5

I like to filter patients by various combination and criteria.

I made a rigid function as follows:

def filter_data(df):
    df = df[
        (df['PREINDEX_N_DRUG_A_TAG']>8) &
        (df['POSTINDEX_N_DRUG_A_TAG']==2) &
        (df['PREINDEX_N_DRUG_B_TAG']==0) &
        (df['POSTINDEX_N_DRUG_B_TAG']>=2)
        ]

However, I want to make the function in such a way that is more flexible, so that the users can specify how ever many of criteria and the form of the criteria.

For example, I would like a function to use a command_dict information such as

command_dict = {
    'Command1': ['PREINDEX_N_DRUG_A_TAG', '>', 8, '&'],
    'Command2': ['POSTINDEX_N_DRUG_A_TAG', '==', 2, '&'],
    'Command3': ['PREINDEX_N_DRUG_B_TAG', '==', 0, '&'],
    'Command4': ['POSTINDEX_N_DRUG_B_TAG', '>=', 2, '&'],
    'Command5': ['PREINDEX_N_DRUG_A_TAG', '>', 8, '&'],
    'Command6': ['PREINDEX_N_DRUG_A_TAG', '>', 8, '&'],
    }

And be able to perform the function as in the rigid function.

Upvotes: 1

Views: 126

Answers (1)

BENY
BENY

Reputation: 323366

This is perfect use case for query 

df.query((''.join([''.join(map(str,x)) for x in command_dict.values()])).strip('&'))

Upvotes: 2

Related Questions