CODEWITHSUNDEEP
pythonpython-3.xpandasdataframe
chessosapiens
chessosapiens

Reputation: 3419

convert SQL full outer join of the same table to Full outer join of pandas dataframe

How can i convert following SQl that do full outer join on a table

FROM df AS df1 FULL OUTER JOIN df AS df2
ON df1.date = df2.date AND df1.product != df2.product

does

     pd.merge(df1,df2, on="date , how="outer")

do the same? i guess not then how can i also tell pandas df1.product != df2.product

Upvotes: 0

Views: 102

Answers (1)

Literal
Literal

Reputation: 827

Unfortunately, it's very hard to do non-equality join predicates in pandas. You'll have to first do your merge, then, assuming the result is stored in variable t, filter you products: t[t['product_x'] != t['product_y']].

There are some packages (pandasql, duckdb) that let you query directly on dataframes, so that you can apply complex predicates, but I heard it can be rather slow. If your dataframe is of an intermediate size (fits comfortably in memory) I guess the first option is the simplest.

Upvotes: 1

Related Questions