Jason Boyd Jason Boyd - 3 months ago 11
Python Question

How do reindex multilevel columns

Version info:

print(sys.version)
3.5.1 |Anaconda 4.1.0 (64-bit)| (default, Jun 15 2016, 15:29:36) [MSC v.1900 64 bit (AMD64)]


I have columns in a data frame that look like this (latitude and longitude are multilevel columns):

+------------+---------------+--------------+--------------+
| CustomerId | StreetAddress | Latitude | Longitude |
+------------+---------------+-------+------+-------+------+
| | count | mean | count | mean |
+----------------------------+-------+------+-------+------+


I would like to get this:

+------------+---------------+-----------+----------+-----------+----------+
| CustomerId | StreetAddress | Lat_count | Lat_mean | Lon_count | Lon_mean |
+------------+---------------+-----------+----------+-----------+----------+


I tried this:

newColumns = ['CustomerId','StreetAddress','Lat_count','Lat_mean','Lon_count','Lon_mean']
data2 = data1.reindex(columns=newColumns)


But that absolutely did not work! I ended up with some kind of crazy multilevel columns with each letter of each string in
newColumns
being a new level.

Update



Here are my columns

data1.columns.to_series()

CustomerId (CustomerId, )
StreetAddress (StreetAddress, )
Latitude count (Latitude, count)
mean (Latitude, mean)
Longitude count (Longitude, count)
mean (Longitude, mean)

Answer

This will do the trick:

data2 = pd.DataFrame(data1.values, columns=newColumns)

And also this:

data1.columns = newColumns
Comments