Pandas check if two columns can be considered the composite key of the dataframe-CodePudding

A sample dataframe:

data = {
    "col_A": ["a","a","b","c"],
    "col_B": [1, 2, 2, 3],
    "col_C": ["demo", "demo", "demo", "demo"]
}

df = pd.DataFrame(data)

Dataframe

col_A   col_B   col_C
a       1       demo
a       2       demo
b       2       demo
c       3       demo

I can easily check if all values in col_A are unique or not by df['col_A'].is_unique. Is there any way to check for two columns i.e. something like df['col_A', 'col_B'].is_unique

If col_A and col_B are the composite key of the data frame or not?

CodePudding user response：

You can set all columns that should be included in the composite key as index and then check for is_unique on the index.

df.set_index(['col_A', 'col_B']).index.is_unique

#True

CodePudding user response：

Sounds like you're just looking for dupes?

df[['col_A', 'col_B']].duplicated().any() # returns False

CodePudding user response：

Use:

df[['col_A', 'col_B']].duplicated().any()