Reputation: 389
I have this pandas dataframe:
open high low close volume
TimeStamp
2016-06-23 10:00:00 586.76 594.00 585.54 589.94 478.176973
2016-06-23 11:00:00 589.94 595.49 588.23 592.63 448.689485
2016-06-23 12:00:00 592.63 592.63 1.50 581.13 527.816891
2016-06-23 13:00:00 581.13 586.33 578.58 580.96 728.424757
As you can see one of the values is not ok. So I want to filter it and change it to the mean of the last 5 values
With this
df['avg']=df['low'].rolling(5).mean().shift()
I get this
open high low close volume avg
TimeStamp
2016-06-23 10:00:00 586.76 594.00 585.54 589.94 478.176973 573.326
2016-06-23 11:00:00 589.94 595.49 588.23 592.63 448.689485 578.438
2016-06-23 12:00:00 592.63 592.63 1.50 581.13 527.816891 583.202
2016-06-23 13:00:00 581.13 586.33 578.58 580.96 728.424757 467.348
And now I want to give to the low the same value of avg. The filter finds value that has a "variance" bigger than 5.
df.loc[(df['high']/df['low'])>5]['low']
open high low close volume avg
TimeStamp
2016-06-23 12:00:00 592.63 592.63 1.5 581.13 527.816891 583.202
But when I try to give the value, it doesn't work..
df.loc[(df['high']/df['low'])>5]['low']=df.loc[(df['high']/df['low'])>5]['avg']
Can you help me?
Upvotes: 2
Views: 96
Reputation: 323316
pandas' dataframe is base on index, so what you need is just
df.loc[(df['high']/df['low'])>5,'low']=df.avg
df
Out[1331]:
open high low close volume avg
0 586.76 594.00 585.54 589.94 478.176973 NaN
1 589.94 595.49 588.23 592.63 448.689485 585.54
2 592.63 592.63 588.23 581.13 527.816891 588.23
3 581.13 586.33 578.58 580.96 728.424757 1.50
Upvotes: 2