I've got a dataframe df1
with date and other values like below:
date value1 value2 value3
20100101 1 2 3
20100102 1 2 3
20100103 1 2 3
20100104 1 3 4
20100105 1 3 4
20100106 1 3 5
20100107 1 3 6
Then I'd like to update some values from another dataframe df2
:
date value1
20100102 2
20100104 3
20100105 4
20100106 5
20100107 6
So the expected outcome will be:
date value1 value2 value3
20100101 1 2 3
20100102 2 2 3
20100103 1 2 3
20100104 3 3 4
20100105 4 3 4
20100106 5 3 5
20100107 6 3 6
As far as I know I can't do this with a left join, is there any fast and easy way to achieve this other than iterate through each date?
Update:
Thanks for all the answers!
I've got another case when df2
has different dates from df1
, eg
date value1
20100102 2
20100104 3
20100105 4
20100106 5
20100107 6
20100108 7
Adding dropna(axis=0, how='any')
to piRSquared's answer will solve this case.
d2.set_index('date').combine_first(
d1.set_index('date')).reset_index().astype(d1.dtypes)
date value1 value2 value3
0 20100101 1 2 3
1 20100102 2 2 3
2 20100103 1 2 3
3 20100104 3 3 4
4 20100105 4 3 4
5 20100106 5 3 5
6 20100107 6 3 6
d1[['date']].merge(d2, 'left').combine_first(d1).astype(d1.dtypes)
date value1 value2 value3
0 20100101 1 2 3
1 20100102 2 2 3
2 20100103 1 2 3
3 20100104 3 3 4
4 20100105 4 3 4
5 20100106 5 3 5
6 20100107 6 3 6
I think this is faster:
In [58]: df.loc[df[df.date.isin(sd.date)].index,'value1'] = sd.value1.values.tolist()
In [59]: df
Out[59]:
date value1 value2 value3
0 20100101 1 2 3
1 20100102 2 2 3
2 20100103 1 2 3
3 20100104 3 3 4
4 20100105 4 3 4
5 20100106 5 3 5
6 20100107 6 3 6
In [61]: %timeit df.loc[df[df.date.isin(sd.date)].index,'value1'] = sd.value1.values.tolist()
1000 loops, best of 3: 703 µs per loop
In [62]: %timeit sd.set_index('date').combine_first(df.set_index('date')).reset_index().astype(df.dtypes)
100 loops, best of 3: 4.08 ms per loop
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.