Python 玩转数据 16 - Pandas 数据处理 追加 df.append()

引言

本文主要介绍 pandas 数据追加 df.append()，更多 Python 进阶系列文章，请参考 Python 进阶学习 玩转数据系列

内容提要：
df1.append()
追加 DataFrame
追加 Series
追加 dict 字典

df1.append()

按行追加拼接，将一个 DataFrame 的行拼接到另外一个 DataFrame 的末尾，返回一个新的 DataFrame。如果列名不在第一个 DataFrame出现，则将以新的列名添加，没有对应内容的会为空。不会改变原来的 DataFrame，只会创建一个新的 DataFrame，包含拼接的数据。

重点：因为会创建一个新的 index 和 data buffer，所以效率不高。推荐用 pd.concat()，而且 pd.concat() 功能更强大，详情请参考 Pandas 数据处理 拼接 pd.concat()

df.append(other: DataFrame | Series[Dtype@append] | Dict[_str, Any], ignore_index: _bool = …, verify_integrity: _bool = …, sort: _bool = …)

参数	说明
other	是它要追加的其他 DataFrame 或者类似序列内容
ignore_index	如果为 True 则重新进行自然索引
verify_integrity	如果为 True 则遇到重复索引内容时报错
sort	进行排序

追加 DataFrame

两个 df 拼接，并保留原索引：

多个 df 拼接，在合并不保留原索引，启用新的自然索引

代码：

import pandas as pd

idnumber = [1,2,5]
fname = ['Kate','John','Eli']
age = [10,20,50]
grade = ['A','B','C']

df1 = pd.DataFrame({'id':idnumber,'fname':fname})
df2 = pd.DataFrame({'age':age,'grade':grade})
df3 = pd.DataFrame({'id':[3,4],'age':[30,40]})
print("df1:\n{}".format(df1))
print("df2:\n{}".format(df2))
print("df3:\n{}".format(df3))

df1_append_df2 = df1.append(df2)
print("df1_append_df2:\n{}".format(df1_append_df2))

df1_append_df2_df3 = df1.append([df2,df3], ignore_index = True)
print("df1_append_df2_df3:\n{}".format(df1_append_df2_df3))

输出：

df1:       
   id fname
0   1  Kate
1   2  John
2   5   Eli
df2:
   age grade
0   10     A
1   20     B
2   50     C
df3:
   id  age
0   3   30
1   4   40
df1_append_df2:
    id fname   age grade
0  1.0  Kate   NaN   NaN
1  2.0  John   NaN   NaN
2  5.0   Eli   NaN   NaN
0  NaN   NaN  10.0     A
1  NaN   NaN  20.0     B
2  NaN   NaN  50.0     C
df1_append_df2_df3:
    id fname   age grade
0  1.0  Kate   NaN   NaN
1  2.0  John   NaN   NaN
2  5.0   Eli   NaN   NaN
3  NaN   NaN  10.0     A
4  NaN   NaN  20.0     B
5  NaN   NaN  50.0     C
6  3.0   NaN  30.0   NaN
7  4.0   NaN  40.0   NaN

追加 Series

拼接 Series 一定要加上参数 ignore_index=True

Series 的 index 会分别对应列名

代码：

import pandas as pd

idnumber = [1,2,5]
fname = ['Kate','John','Eli']

df1 = pd.DataFrame({'id':idnumber,'fname':fname})
s1 = pd.Series(["aa", "bb"], index=["a", "b"])
s2 = pd.Series(["110", "Supermen"], index=["id", "fname"])

print("df1:\n{}".format(df1))
print("s1:\n{}".format(s1))
print("s2:\n{}".format(s2))

df1_appened_s1 = df1.append(s1, ignore_index=True)
df1_appened_s2 = df1.append(s2, ignore_index=True)

print("df1_appened_s1:\n{}".format(df1_appened_s1))
print("df1_appened_s2:\n{}".format(df1_appened_s2))

输出：

df1:
   id fname
0   1  Kate
1   2  John
2   5   Eli
s1:
a    aa
b    bb
dtype: object
s2:
id            110
fname    Supermen
dtype: object
df1_appened_s1:
    id fname    a    b
0  1.0  Kate  NaN  NaN
1  2.0  John  NaN  NaN
2  5.0   Eli  NaN  NaN
3  NaN   NaN   aa   bb
df1_appened_s2:
    id     fname
0    1      Kate
1    2      John
2    5       Eli
3  110  Supermen

追加 dict 字典

代码：

import pandas as pd

idnumber = [1,2,5]
fname = ['Kate','John','Eli']

df1 = pd.DataFrame({'id':idnumber,'fname':fname})
dicts_1 =[{"a":"aa", "b":"bb"},{"a":"aaa", "b":"bbb"}]
dicts_2 =[{"id":"110", "fname":"Supermen"},{"id":"111", "fname":"Superwoman"}]


print("df1:\n{}".format(df1))
print("dicts_1:\n{}".format(dicts_1))
print("dicts_1:\n{}".format(dicts_1))

df1_appened_dicts_1 = df1.append(dicts_1)
df1_appened_dicts_2 = df1.append(dicts_2, ignore_index=True)

print("df1_appened_dicts_1:\n{}".format(df1_appened_dicts_1))
print("df1_appened_dicts_2:\n{}".format(df1_appened_dicts_2))

输出：

df1:
   id fname
0   1  Kate
1   2  John
2   5   Eli
dicts_1:
[{'a': 'aa', 'b': 'bb'}, {'a': 'aaa', 'b': 'bbb'}]
dicts_1:
[{'a': 'aa', 'b': 'bb'}, {'a': 'aaa', 'b': 'bbb'}]
df1_appened_dicts_1:
    id fname    a    b
0  1.0  Kate  NaN  NaN
1  2.0  John  NaN  NaN
2  5.0   Eli  NaN  NaN
0  NaN   NaN   aa   bb
1  NaN   NaN  aaa  bbb
df1_appened_dicts_2:
    id       fname
0    1        Kate
1    2        John
2    5         Eli
3  110    Supermen
4  111  Superwoman