python - How I do find median using pandas on a dataset?

Question

I have dataframe data which has 3 columns - Date, segment and metric. I am doing the following:

data = pandas.read_csv("Filename.csv")
ave = data.groupby('Segment').mean() #works
ave = data.groupby('Segment').median() #gives error
ave['median'] = data.groupby('Segment').median()

Traceback (most recent call last):
  File "<input>", line 1, in <module>
  File "/usr/lib/pymodules/python2.7/pandas/core/frame.py", line 1453, in __setitem__
    self._set_item(key, value)
  File "/usr/lib/pymodules/python2.7/pandas/core/frame.py", line 1488, in _set_item
    NDFrame._set_item(self, key, value)
  File "/usr/lib/pymodules/python2.7/pandas/core/generic.py", line 301, in _set_item
    self._data.set(key, value)
  File "/usr/lib/pymodules/python2.7/pandas/core/internals.py", line 616, in set
    assert(value.shape[1:] == self.shape[1:])
AssertionError

Answer 1

What error do you get with?

ave = data.groupby('Segment').median()

I think that should work, maybe there's something in your data causing the error, like nan's, im just guessing. You could try applying your own median function to see if you can work around the cause of the error, something like:

def mymed(group):
    return np.median(group.dropna())

ave = data.groupby('segment')['Metric'].apply(mymed)

It would be easier if you could provide some sample data which replicates the error.

Here is a different approach, you can add the median back to your original dataframe, the median for the metric column becomes:

data['metric_median'] = data.groupby('Segment')['Metric'].transform('median')

Wether its useful to have the median of the group attached to each datapoint depends a bit what you want to do afterwards.