I am pretty new to python/pandas sorry if this is a simple question. I currently have 2 dataframes, one with a date range and another with product SKUs. I want to create a data frame that repeats all SKUs for each date.
I am currently doing this by using iterrows() but it is terribly inefficient for large datasets.
Thanks in advance.
Example:
dates_df:
Date
0 2016-01-01
1 2016-01-02
2 2016-01-03
sku_df:
SKU
0 001
1 002
2 003
result_df:
Date SKU
2016-01-01 001
2016-01-01 002
2016-01-01 003
2016-01-02 001
2016-01-02 002
2016-01-02 003
2016-01-03 001
2016-01-03 002
2016-01-03 003
you can use itertools.product :
In [30]: from itertools import product
In [31]: pd.DataFrame(list(product(dates.Date, sku.SKU)), columns=['Date','SKU'])
Out[31]:
Date SKU
0 2016-01-01 001
1 2016-01-01 002
2 2016-01-01 003
3 2016-01-02 001
4 2016-01-02 002
5 2016-01-02 003
6 2016-01-03 001
7 2016-01-03 002
8 2016-01-03 003
or using Pandas's cartesian product:
In [136]: pd.merge(dates.assign(key='x'), sku.assign(key='x'), on='key').drop('key', 1)
Out[136]:
Date SKU
0 2016-01-01 001
1 2016-01-01 002
2 2016-01-01 003
3 2016-01-02 001
4 2016-01-02 002
5 2016-01-02 003
6 2016-01-03 001
7 2016-01-03 002
8 2016-01-03 003
Source DFs:
In [32]: dates
Out[32]:
Date
0 2016-01-01
1 2016-01-02
2 2016-01-03
In [33]: sku
Out[33]:
SKU
0 001
1 002
2 003
Timing for two DFs 1K rows each, so the resulting DF will have 1M rows:
In [153]: dates = pd.DataFrame({'Date':pd.date_range('2000-01-01', periods=1000)})
...: sku = pd.DataFrame({'SKU':np.arange(1, 1001).astype(str)})
...: sku.SKU = sku.SKU.str.zfill(3)
...:
In [154]: dates.shape
Out[154]: (1000, 1)
In [155]: sku.shape
Out[155]: (1000, 1)
In [156]: %timeit pd.DataFrame(list(product(dates.Date, sku.SKU)), columns=['Date','SKU'])
1 loop, best of 3: 667 ms per loop
In [157]: %timeit pd.merge(dates.assign(key='x'), sku.assign(key='x'), on='key').drop('key', 1)
1 loop, best of 3: 222 ms per loop
In [158]: len(pd.DataFrame(list(product(dates.Date, sku.SKU)), columns=['Date','SKU']))
Out[158]: 1000000
In [159]: len(pd.merge(dates.assign(key='x'), sku.assign(key='x'), on='key').drop('key', 1))
Out[159]: 1000000
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.