Pandas equivalent of SQL non-equi JOIN

Question

So I've got 2 data-frames I'd like to merge together.

I'm merging on 3 columns, 2 is an easy join.

joined_df = pd.merge(df1, df2, how='left', on=['name', 'city'])

I want this to be using a third column, but it's going to be a comparison, something like this:

joined_df = pd.merge(df1, df2, how='left',
on=['name', 'city', 'df1.year' >= 'df2.year_min'])

Not sure what the right syntax is here.

If it was SQL, it would be easy for me.

SELECT * FROM df1
JOIN df2 on (df1.name = df2.name and df1.year = df2.year and df1.year > df2.year_min)

Any assistance?

Answer 1

Pandas merge only supports equi-joins. You'll need to add a second step that filters the result, something like this:

joined_df = df1.merge(df2, how='left', on=['name', 'city'])
joined_df = joined_df[joined_df.year > joined_df.year_min]

Answer 2

You can using merge_asof, default is backward merge

pd.merge_asof(df1,df2, left_on='year',right_on='joined_df', by=['name', 'city'])