How to use dynamic string to filter data frame using Python Pandas

Question

DataFrame

    PROJECT  CLUSTER_x  MARKET_x  CLUSTER_y  MARKET_y     Exist
0   P17      A          CHINA     C          CHINA        both
1   P18      P          INDIA     P          INDIA        both
2   P16      P          AMERICA   P          AMERICA      both
3   P19      P          INDIA     P          JAPAN        both

This below code works perfectly alright and gives output as index 0 and 3

df_mismatched = df_common[ (df_common['MARKET_x'] != df_common['MARKET_y']) | (df_common['CLUSTER_x'] != df_common['CLUSTER_y']) ]

How we can dynamlically build such filter criteria? something like below code, so that next time hardcoding won't be necessary

str_common = '(df_common["MARKET_x"] != df_common["MARKET_y"]) | (df_common["CLUSTER_x"] != df_common["CLUSTER_y"])'
df_mismatched = df_common[str_common]

Answer 1

For the dynamic purpose, you can use query in python like:

con = "(MARKET_x!=MARKET_y)|(CLUSTER_x!=CLUSTER_y)"
print(df.query(con))

  PROJECT CLUSTER_x MARKET_x CLUSTER_y MARKET_y Exist
0     P17         A    CHINA         C    CHINA  both
3     P18         P    INDIA         P    JAPAN  both

Remember that if the columns names have spaces or special characters it fails to produce the right results.