If the index is not unique, then using join to combine the two dataframes creates all possible combinations of matching indexes. pd.concat does the expected here, which is a 1:1 mapping.
Since the index is based on the monomers hash, it's not unique across monomers with different distributions. In Mark's case, what was supposed to be 170 rows turns into 2270 rows.
If the index is not unique, then using
join
to combine the two dataframes creates all possible combinations of matching indexes.pd.concat
does the expected here, which is a 1:1 mapping.Since the index is based on the monomers hash, it's not unique across monomers with different distributions. In Mark's case, what was supposed to be 170 rows turns into 2270 rows.