Pandas Tutorials | Projects | Interview Questions Tutorial

Left-Right and Outer Merge

❮ Previous Next ❯

Till now, we have been dealing with inner merge which, only includes matching rows. If rows don't match perfectly, we lose the unmatched rows.

Let's check our dataset first

import pandas as pd

store_a = pd.read_csv('store_a.csv')
store_b = pd.read_csv('store_b.csv')
print(store_a.head())
print(store_b.head())

Now, if we want to use an outer join which would include all rows from both the tables, and if they don't match it will be filled with None or nan (not a number). We'll use the parameter 'how' and passed 'outer' as a value.
```
outer = pd.merge(store_a, store_b, how='outer')
print(outer.head())
```

Left Merge includes all rows from the first (left) table, but only rows from the second (right) table that match the first table.
It matters how the arguments are presented. Only rows that occur in Table A will be returned if the initial DataFrame is Table A and a left join is performed.
By placing Table A first, we receive all of Table A's rows and only those clients of Table B who are also Rows of Table A.
```
store_a_b_left = pd.merge(store_a, store_b, how = 'left')
```
```
store_b_a_left = pd.merge(store_b, store_a, how = 'left')
```
Here, these two tables will be different as store_a has some rows which is not present in store_b and vice-versa.

Right merge is the opposite of left merge.
Here, the merged table will include all rows from the second (right) table, but only rows from the first (left) table that match the second table.
```
store_a_b_right = pd.merge(store_a, store_b, how = 'right')
store_b_a_right = pd.merge(store_b, store_a, how = 'right')
```

❮ Previous Next ❯