CODEWITHSUNDEEP
pythonpandastime-series
Ivan T
Ivan T

Reputation: 1046

Align numeric indexes to temporal period in pandas

I have a data frame like this:

0 0.131567 1 0.098878 2 0.102000 . ... 94574 0.088912 94575 0.089399

and I know it corresponds to some time interval eg. 15:57:32 - 16:50:05. So I would like to convert indexes from sequence of number to sequence of timestamps. So that output will look like this 

 15:57:32.000000   0.131567  
 15:57:32.031000   0.102000  
             ...   ...
 16:50:04.969000   0.088912
 16:50:04.000000   0.089399`

Perhaps I can do it with the following code:

>     rng = date_range('15:57:32.000000','16:50:04.000000' , freq=??) 
>     ts = Series(data, index=rng)

But how should I choose frequency then? Are there any other way to do it or I am on the right path?

Upvotes: 0

Views: 135

Answers (1)

B. M.
B. M.

Reputation: 18658

you can specify a multiplier in the freq field. help(pd.date_range) gives :

freq : string or DateOffset, default 'D' (calendar daily) Frequency strings can have multiples, e.g. '5H'

rng = pd.date_range('15:57:32','16:50:04',freq='33328us')

Then len(rng) is exactly 94576 :

2016-03-15 15:57:32         0.881135
2016-03-15 15:57:32.033328  0.237618
2016-03-15 15:57:32.066656  0.761269
....
2016-03-15 16:50:03.962272  0.820865
2016-03-15 16:50:03.995600  0.181762

Upvotes: 1

Related Questions