如何根据两个数据框中的 ID 信息查找值？

Question

包含订单信息的第一个数据帧。一个 Lead Order 可能有多个 orderid。 另一个dataframe有一个OrderID列表，想用dataframe1作为参考来查找LeadOrderID，请问如何使用python（Pandas）查找LeadOrderID？ 谢谢你的帮助。真的很感激。

Answer 1

您应该将panda.merge() 与on=['OrderID'] 和how='inner' 一起使用。

In [207]: df1 = pd.DataFrame({'OrderID':[i for i in range(10)], 'Lead Order':[1,3,5,8,6,7,7,5,2,1]}, index=[0,1,2,3,4,5,6,7,8,9])

In [208]: df1
Out[208]: 
   OrderID  Lead Order
0        0           1
1        1           3
2        2           5
3        3           8
4        4           6
5        5           7
6        6           7
7        7           5
8        8           2
9        9           1

In [209]: df2 = pd.DataFrame({'OrderID':[3,8,6,2]}, index=[0,1,2,3])

In [210]: df2
Out[210]: 
   OrderID
0        3
1        8
2        6
3        2

In [211]: df3 = pd.merge(df1, df2, on=['OrderID'], how='inner')

In [212]: df3
Out[212]: 
   OrderID  Lead Order
0        2           5
1        3           8
2        6           7
3        8           2

Answer 2

这个答案包括处理多个问题 OrderID(s) 列的行内的值。

没有注释的完整代码在下面

# imports
import pandas as pd
import numpy as np

# create sample dataframe
df_orig = \
    pd.DataFrame({'OrderID(s)':['0001, 0007, 0002', '0008', '0009, 0005, 0003',],
                  'Lead Order': ['00011', '00022', '00033']})

df_orig

          OrderID(s)    Lead Order
0   0001, 0007, 0002    00011
1               0008    00022
2   0009, 0005, 0003    00033

-

# force df values to strings
# this makes splitting of multiple
# values in OrderID(s) easier
df_orig = df_orig.astype(str)

# series created from data within df_orig['OrderID(s)'] column
# remove spaces and split by commas
split_col = df_orig['OrderID(s)'].str.replace(' ', '').str.split(",")
print(split_col)

0    [0001, 0007, 0002]
1                [0008]
2    [0009, 0005, 0003]
Name: OrderID(s), dtype: object

-

# find length of each split_col row (how many OrderIDs in each row).
# these values will be used to duplicate rows in the
# df_orig dataframe with the numpy repeat function
repeats = split_col.str.len().values
print(repeats)

[3 1 3]

-

# concatenate all values in orderid_column.
# the length of this array will be the same as the length
# of the df_stack_ids dataframe
orderid_col = np.concatenate(split_col.values)
print(orderid_col)

['0001' '0007' '0002' '0008' '0009' '0005' '0003']

-

# use pandas iloc and numpy repeat function to make a dataframe with
# rows from df_orig duplicated according to the number of
# df_orig['OrderID(s)'] values in each row relating to a common
#Lead Order value (using repeats input from above)
df_stack_ids = df_orig.iloc[np.repeat(df_orig.index.values, repeats)]. \
    reset_index(drop=True)

df_stack_ids

          OrderID(s)    Lead Order
0   0001, 0007, 0002    00011
1   0001, 0007, 0002    00011
2   0001, 0007, 0002    00011
3               0008    00022
4   0009, 0005, 0003    00033
5   0009, 0005, 0003    00033
6   0009, 0005, 0003    00033

-

# add the orderid_col to dataframe
df_stack_ids['OrderID'] = orderid_col

df_stack_ids

          OrderID(s)    Lead Order  OrderID
0   0001, 0007, 0002         00011     0001
1   0001, 0007, 0002         00011     0007
2   0001, 0007, 0002         00011     0002
3               0008         00022     0008
4   0009, 0005, 0003         00033     0009
5   0009, 0005, 0003         00033     0005
6   0009, 0005, 0003         00033     0003

-

# get rid of the original OrderID(s) column
df_stack_ids = df_stack_ids[['OrderID', 'Lead Order']]


# this may be enough to answer the question
# because each order id has a corresponding
# lead order

df_stack_ids

    OrderID Lead Order
0      0001      00011
1      0007      00011
2      0002      00011
3      0008      00022
4      0009      00033
5      0005      00033
6      0003      00033

-

# to find matches for a specific list of order ids,
# continue...
# sort the OrderID column for easy reference and
# reset index
df_stack_ids = df_stack_ids.sort_values(by=['OrderID'])
df_stack_ids.index = range(len(df_stack_ids))


# create sample dataframe with a few order ids for lookup
df_find_lead = pd.DataFrame({'OrderID': ['0001', '0002', '0005']})
# force to string type for matching with df_stack_ids values
# when merging
df_find_lead = df_find_lead.astype(str)

df_find_lead

    OrderID
0      0001
1      0002
2      0005

-

# merge values from df_stack_ids['Lead Order'] column
df_found_lead = pd.merge(df_find_lead, df_stack_ids,
                         on=['OrderID'], how='inner')

df_found_lead

    OrderID Lead Order
0      0001      00011
1      0002      00011
2      0005      00033

-

# if all original order data is formatted as numbers,
# convert result dataframe back to integers
df_found_lead.astype(int)

    OrderID Lead Order
0         1         11
1         2         11
2         5         33

完整代码：

import pandas as pd
import numpy as np

df_orig = \
    pd.DataFrame({'OrderID(s)':['0001, 0007, 0002', '0008', '0009, 0005, 0003',],
                  'Lead Order': ['00011', '00022', '00033']})

df_orig = df_orig.astype(str)
split_col = df_orig['OrderID(s)'].str.replace(' ', '').str.split(",")

repeats = split_col.str.len().values
orderid_col = np.concatenate(split_col.values)

df_stack_ids = df_orig.iloc[np.repeat(df_orig.index.values, repeats)]. \
    reset_index(drop=True)

df_stack_ids['OrderID'] = orderid_col
df_stack_ids = df_stack_ids[['OrderID', 'Lead Order']]
df_stack_ids = df_stack_ids.sort_values(by=['OrderID'])
df_stack_ids.index = range(len(df_stack_ids))

df_find_lead = pd.DataFrame({'OrderID': ['0001', '0002', '0005']})
df_find_lead = df_find_lead.astype(str)

df_found_lead = pd.merge(df_find_lead, df_stack_ids, on=['OrderID'], how='inner')
df_found_lead.astype(int)