Menu
I currently have a DataFrame laid out as:
Jan Feb Mar Apr ... 2001 1 12 12 19 2002 9 ... 2003 ... and I would like to "unpivot" the data to look like:
Date Value Jan 2001 1 Feb 2001 1 Mar 2001 12 ... Jan 2002 9 What is the best way to accomplish this using pandas/NumPy?
668
In pandas, you can use the melt() function to unpivot a DataFrame – converting it from a wide format to a long format. This function uses the following basic syntax: df_unpivot = pd. melt(df, id_vars='col1', value_vars=['col2', 'col3', ...])
We can use pivot() function to unmelt a DataFrame object and get the original dataframe. The pivot() function 'index' parameter value should be same as the 'id_vars' value. The 'columns' value should be passed as the name of the 'variable' column. The unmelted DataFrame values are the same as the original DataFrame.
Pd. melt allows you to 'unpivot' data from a 'wide format' into a 'long format', perfect for my task taking 'wide format' economic data with each column representing a year, and turning it into 'long format' data with each row representing a data point.
An unpivot transformation is one way to transform data from a short/wide to a tall/skinny format. When the data types of source columns differ, the varying data is converted to a common data type so the source data can be part of one single column in the new data set.
You just have to do df.unstack() and that will create a MultiIndexed Series with month as a first level and the year as the second level index. If you want them to be columns then just call reset_index() after that.
>>> df Jan Feb 2001 3 4 2002 2 7 >>> df.unstack() Jan 2001 3 2002 2 Feb 2001 4 2002 7 >>> df = df.unstack().reset_index(name='value') >>> df level_0 level_1 value 0 Jan 2001 3 1 Jan 2002 2 2 Feb 2001 4 3 Feb 2002 7 >>> df.rename(columns={'level_0': 'month', 'level_1': 'year'}, inplace=True) >>> df month year value 0 Jan 2001 3 1 Jan 2002 2 2 Feb 2001 4 3 Feb 2002 7 If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With
