Pandas tricks — Indexing

Indexes in pandas dataframe are a very important aspect of data manipulations, as they are the best placeholders that can be used to do various data wrangling activities.

With this post, I will try to explain index in pandas data frames, how to deal with them and few use cases where they can be useful.

First of all, indexes are nothing but the row location in the data frame, by default it starts with 0.

Lets read a random csv file.

So as you can see, an index is by default provided to the dataframe, starting with 0.

If we want to start the index with 1 -

If we want to see the index details -

To store the index of the dataframe in a list for further iterative processings -

To make the index as one of the dataframe columns -

with reset_index(), you can see that the old indexes became the one of the columns, and the dataframe is by default provided new set of index, again starting with 0.

But if we just want to reset the index starting with 0 again, without getting a new column, use (drop=True) argument in reset_index() -

To set one of the columns as the index -

One use case why would we like to set a column as an index is, lets say we set the VALUE column as the index -

…and now we want to slice the dataframe to extract all the records that has value = 4.7, then we can use loc() method -

Still if we want to use the original index to process the dataframe, use iloc() method instead -

To manually set the index -

To create a new set of index, use reindex() -

In the new set of index, if the dataframe has similar index name, then the values are retained, and new indexes will be assigned NaN.

Git repo -https://github.com/shekharkhandelwal1983/pandas/blob/master/pandas_indexing.ipynb