Home > Article > Backend Development > How to Merge Two DataFrames Based on Index in Pandas?
When working with dataframes, it is often necessary to combine them based on specific criteria. In this case, the objective is to merge two dataframes, df1 and df2, by index.
By default, the merge() function in Python's Pandas library expects column-based matching. However, merging on index is possible using specific parameters.
To perform an inner join, where only rows with matching indices are retained, use the following code:
<code class="python">pd.merge(df1, df2, left_index=True, right_index=True)</code>
This operation produces the following output:
id | begin | conditional | confidence | discoveryTechnique | concept |
---|---|---|---|---|---|
278 | 56 | false | 0.00 | 1 | A |
421 | 18 | false | 0.00 | 1 | B |
Alternatively, a left join can be performed using the join() method:
<code class="python">df1.join(df2)</code>
This results in:
id | begin | conditional | confidence | discoveryTechnique | concept |
---|---|---|---|---|---|
278 | 56 | false | 0.00 | 1 | NaN |
421 | 18 | false | 0.00 | 1 | B |
2 | 56 | false | 0.00 | 1 | NaN |
5 | 37 | false | 0.20 | 1 | NaN |
Finally, an outer join can be achieved using the concat() function:
<code class="python">pd.merge(df1, df2, left_index=True, right_index=True)</code>
The resulting dataframe includes all rows from both input dataframes:
id | begin | conditional | confidence | discoveryTechnique | concept |
---|---|---|---|---|---|
278 | 56 | false | 0.00 | 1 | A |
421 | 18 | false | 0.00 | 1 | B |
2 | 56 | false | 0.00 | 1 | NaN |
5 | 37 | false | 0.20 | 1 | NaN |
8 | 36 | false | 0.50 | 1 | NaN |
NaN | 37 | false | 0.30 | 2 | NaN |
Remember that merging on the index is not a common practice and should be considered carefully to avoid data loss or integrity issues. If merging by index is unavoidable, the provided methods offer flexible options to achieve the desired results.
The above is the detailed content of How to Merge Two DataFrames Based on Index in Pandas?. For more information, please follow other related articles on the PHP Chinese website!