converting dataframe columns full of minutes and seconds.

[ryan h]

So I have a dataframe like this.   df = pd.DataFrame({'A':[u'23:30', u'31:42']})

How do I convert them to floats so the output is like this A        # divide the seconds by 60 to convert the seconds into decimals.  

                                                           23.5

                                                           31.7                   

I cant find anything in the book about converting just minutes and seconds.  I only see stuff on dates.  

[Wouter Overmeire]

2013/5/14 ryan h <ryan...@gmail.com>

--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+un...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

In [34]: df = pd.DataFrame({'A':[u'23:30', u'31:42']})

In [35]: df

Out[35]:

       A

0  23:30

1  31:42

In [36]: def convert(x):

   ....:     m, s = x.split(':')

   ....:     return float(m) + float(s) / 60.0

   ....:

In [37]: convert('23:30')

Out[37]: 23.5

In [38]: df.applymap(convert)

Out[38]:

      A

0  23.5

1  31.7

[ryan h]

It works, but the actual frame as other columns that consists of other data so It won't work for it.

in[21]: df = pd.DataFrame({'A':[u'23:30', u'31:42'],

                                        'B':[25.3,456]})

in[22]: df

 

out[22] : A       B

           23:30  25.3

           31:42  456.0

in[23]: df.applymap(convert) gives this error AttributeError: ("'numpy.float64' object has no attribute 'split'", u'occurred at index B')

and I can't just convert the specific column. 

in[24]" df['A'].applymap(convert) AttributeError: 'Series' object has no attribute 'applymap'

I tried: in

in[24]: colA = df['A']

in[25]: colA.applymap(convert)

but get this error

AttributeError: 'Series' object has no attribute 'applymap'

[ryan h]

I guess I just need to change the function so that it leaves the float data alone.  

[Wouter Overmeire]

2013/5/15 ryan h <ryan...@gmail.com>

--
You received this message because you are subscribed to the Google Groups "PyData" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pydata+un...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

On a Series apply (iso applymap) needs to be used to process element by element.

In [6]: df['A'] = df['A'].apply(convert)

In [7]: df

Out[7]:

      A      B

0  23.5   25.3

1  31.7  456.0

Or you select out the columns in such a way that always a DataFrame is returned and applymap can be used. By using the list as indexer, multiple columns can be converted in one go

In [21]: df[['A']] = df[['A']].applymap(convert)

In [22]: df

Out[22]:

      A      B

0  23.5   25.3

1  31.7  456.0

Example for converting multiple columns (and leaving other columns untouched):

In [23]: df = pd.DataFrame({'A': [u'23:30', u'31:42'],

   ....:                    'B': [u'11:10', u'8:23'],

   ....:                    'C': ['foo', 'bar']}

   ....: )

In [24]: df

Out[24]:

       A      B    C

0  23:30  11:10  foo

1  31:42   8:23  bar

In [25]: cols = ['A', 'B']

In [26]: df[cols] = df[cols].applymap(convert)

In [27]: df

Out[27]:

      A          B    C

0  23.5  11.166667  foo

1  31.7   8.383333  bar

[ryan h]

That's what I needed.  Thank You.