CODEWITHSUNDEEP
pythonpandas
user3316115
user3316115

Reputation: 75

Pandas - diagonal / shift

Looking for advice please.

In my DF below, I would like to subract the 'difference' column value (red square), from the value in 'trailing sl' column (blue square), but shifted to lag .shift(1).

So the new values would be:

1.08778 - 0.00115
1.08663 - 0.00077
1.08586 - 0.00059
etc

I've tried .loc with a .shift value, but that's not quite right.

df.loc[df.difference != np.nan, 'trailing_sl'] = df.trailing_sl.shift(1) - df.difference

I'm stuck tbh.

enter image description here

Any advice on a route forward?

Upvotes: 0

Views: 129

Answers (2)

user3316115
user3316115

Reputation: 75

FWIW, my answer was to make a cumulative sum series of the 'difference' column, and then subract the 'trailing_sl' value from the new series.

It seems to work, so far at least :)

Upvotes: 0

jezrael
jezrael

Reputation: 862591

Use Series.notna:

m = df['difference'].notna()
df.loc[m, 'trailing_sl'] = df.trailing_sl.shift() - df['difference']

Or:

df['trailing_sl'] = df.trailing_sl.shift().sub(df['difference']).fillna(df['trailing_sl'])

Upvotes: 2

Related Questions