How to replace Values in Data Frame

[erik...@hotmail.com]

Hi

i have a df like this, and going to perlace the missing values -9999
to NaN or whatever
how can i do it easy way?

Thank's a lot;-)

given:

In [215]: df1
Out[215]:
data1 data2
2005-01-31 -9999 -9999
2005-02-28 -9999 -9999
2005-03-31 2 22
2005-04-30 3 21
2005-05-31 4 20
2005-06-30 5 19
2005-07-31 6 18
2005-08-31 7 17
2005-09-30 8 16
2005-10-31 9 15
2005-11-30 10 14
2005-12-31 11 13

desired result:
In [215]: df1
Out[215]:
data1 data2
2005-01-31 NaN NaN
2005-02-28 NaN NaN
2005-03-31 2 22
2005-04-30 3 21
2005-05-31 4 20
2005-06-30 5 19
2005-07-31 6 18
2005-08-31 7 17
2005-09-30 8 16
2005-10-31 9 15
2005-11-30 10 14
2005-12-31 11 13

[Adam Klein]

Hey Erik, 

You can always just loop through:

for col in df1:

    for i, val in enumerate(col):

        if df[col][i] == -9999:

            df[col][i] = numpy.NaN

[Adam Klein]

Of course, as soon as I wrote it I realized there's a better way:

for k, v in df1.iteritems():

    v[v == -9999] = numpy.NaN

[Wes McKinney]

df[df == -9999] = np.nan 

should work if the DataFrame is homogeneous (all columns floating point)

[erik...@hotmail.com]

Thanks for the super quick responses, pandas is just great:-)

On 9 Nov., 17:09, Wes McKinney <wesmck...@gmail.com> wrote:
> On Nov 9, 2011, at 11:01 AM, Adam Klein <a...@lambdafoundry.com> wrote:
>
>
>
>
>
>
>
>
>
> On Wed, Nov 9, 2011 at 10:58 AM, Adam Klein <a...@lambdafoundry.com> wrote: