df.melt

pandas.melt(frame, id_vars=None, value_vars=None, var_name=None, value_name=’value’, col_level=None, ignore_index=True)[source]
Unpivot a DataFrame from wide to long format, optionally leaving identifiers set.
This function is useful to massage a DataFrame into a format where one or more columns are identifier variables (idvars), while all other columns, considered measured variables (value_vars), are “unpivoted” to the row axis, leaving just two non-identifier columns, ‘variable’ and ‘value’.
Parameters
**id_vars_tuple, list, or ndarray, optional
Column(s) to use as identifier variables.
valuevars_tuple, list, or ndarray, optional
Column(s) to unpivot. If not specified, uses all columns that are not set as id_vars.
varname_scalar
Name to use for the ‘variable’ column. If None it uses frame.columns.name or ‘variable’.
valuename_scalar, default ‘value’
Name to use for the ‘value’ column.
collevel_int or str, optional
If columns are a MultiIndex then use this level to melt.
ignoreindex_bool, default True**
If True, original index is ignored. If False, the original index is retained. Index labels will be repeated as necessary.

Returns: DataFrame

image.png
例子：

  1. >>> df
  2. ...
  3. ...
  4. >>>df
  5.     Country Name    1990    2015
  6. 40    China    73.558    44.500
  7. 55    Germany    26.882    22.800
  8. 119    Japan    22.661    8.619
  9. 251    United States    24.700    18.329
  1. # melt year columns  and convert year to date time
  2. df_melt = df.melt(id_vars='Country Name', value_vars = ['1990', '2015'])
  3. df_melt.columns = ['country','year', 'variable']
  4. df_melt['year'] = df_melt['year'].astype('datetime64[ns]').dt.year
  1. >>> df_melt
  2. country    year    variable
  3. 0    China    1990    73.558
  4. 1    Germany    1990    26.882
  5. 2    Japan    1990    22.661
  6. 3    United States    1990    24.700
  7. 4    China    2015    44.500
  8. 5    Germany    2015    22.800
  9. 6    Japan    2015    8.619
  10. 7    United States    2015    18.329