我有一個(gè)看起來像這樣的熊貓數(shù)據(jù)框：我想把它轉(zhuǎn)換成這種格式：在 Pandas 中執(zhí)行此操作的合適技術(shù)是什么？要在第一張圖片中生成原始數(shù)據(jù)幀，您可以使用以下數(shù)據(jù)。使用此數(shù)據(jù)生成數(shù)據(jù)框import pandas as pddf=pd.DataFrame.from_dict(data_dict)data_dict = {'female_count': {0: 65.0, 1: 13.0, 2: 51.0, 3: 19.0, 4: 53.0, 5: 146.0}, 'jobTitle': {0: 'Associate', 1: 'Intern', 2: 'Key Holder', 3: 'Retail Store Manager', 4: 'Seasonal Sales Associate', 5: 'other_jobTitles'}, 'male_count': {0: 44.0, 1: 17.0, 2: 32.0, 3: 6.0, 4: 26.0, 5: 125.0}, 'mean_z_score_female': {0: 0.0, 1: 0.0, 2: -0.352017669928334, 3: 0.08210979257229986, 4: -0.10918109741978921, 5: 0.23156942728286573}, 'mean_z_score_male': {0: -0.04759213253110343, 1: 0.0, 2: -0.28872649090170965, 3: -0.0027563646918771063, 4: 0.0, 5: 0.6133138806146933}}
5 回答
肥皂起泡泡
TA貢獻(xiàn)1829條經(jīng)驗(yàn) 獲得超6個(gè)贊
這里有一個(gè)使用 `concat' 的相當(dāng)簡(jiǎn)單的方法：
f = df[["jobTitle", "female_count", "mean_z_score_female"]].rename(columns = {"female_count": "count",
"mean_z_score_female": "mean_z_score"})\
.assign(gender="female")
m = df[["jobTitle", "male_count", "mean_z_score_male"]].rename(columns = {"male_count": "count",
"mean_z_score_male": "mean_z_score"})\
.assign(gender="male")
pd.concat([m, f]).sort_values("jobTitle")
輸出是：
jobTitle count mean_z_score gender
0 Associate 44.0 -0.047592 male
0 Associate 65.0 0.000000 female
1 Intern 17.0 0.000000 male
1 Intern 13.0 0.000000 female
2 Key Holder 32.0 -0.288726 male
2 Key Holder 51.0 -0.352018 female
3 Retail Store Manager 6.0 -0.002756 male
3 Retail Store Manager 19.0 0.082110 female
4 Seasonal Sales Associate 26.0 0.000000 male
4 Seasonal Sales Associate 53.0 -0.109181 female
5 other_jobTitles 125.0 0.613314 male
5 other_jobTitles 146.0 0.231569 female
幕布斯7119047
TA貢獻(xiàn)1794條經(jīng)驗(yàn) 獲得超8個(gè)贊
這是的工作pd.wide_to_long，但您首先必須重命名某些列，即female_countandmale_count和：count_femalecount_male
df.columns = ["_".join(entry.split("_")[::-1])
if "count" in entry else entry
for entry in df]
]
print(df.columns)
Index(['count_female', 'jobTitle', 'count_male', 'mean_z_score_female',
'mean_z_score_male'],
dtype='object')
print (pd.wide_to_long(df, stubnames=["count","mean_z_score"],
i="jobTitle",j="gender", sep="_", suffix="\w+"))
count mean_z_score
jobTitle gender
Associate female 65.0 0.000000
Intern female 13.0 0.000000
Key Holder female 51.0 -0.352018
Retail Store Manager female 19.0 0.082110
Seasonal Sales Associate female 53.0 -0.109181
other_jobTitles female 146.0 0.231569
Associate male 44.0 -0.047592
Intern male 17.0 0.000000
Key Holder male 32.0 -0.288726
Retail Store Manager male 6.0 -0.002756
Seasonal Sales Associate male 26.0 0.000000
other_jobTitles male 125.0 0.613314
慕妹3146593
TA貢獻(xiàn)1820條經(jīng)驗(yàn) 獲得超9個(gè)贊
這是一個(gè)與 Roy2012 非常相似的答案，它使用append：
df_new = None
for gender in ['male','female']:
df_gender = (df[['jobTitle',f'{gender}_count',f'mean_z_score_{gender}']]
.rename(columns = {f'{gender}_count':'count',
f'mean_z_score_{gender}': 'mean_z_score'}))
df_gender['gender'] = gender
df_new = df_gender if df_new is None else df_new.append(df_gender)
df_new = df_new.sort_values(by=['jobTitle','gender'],
axis=0).reset_index(drop=True)
print(df_new)
輸出是：
jobTitle count mean_z_score gender
0 Associate 65.0 0.000000 female
1 Associate 44.0 -0.047592 male
2 Intern 13.0 0.000000 female
3 Intern 17.0 0.000000 male
4 Key Holder 51.0 -0.352018 female
5 Key Holder 32.0 -0.288726 male
6 Retail Store Manager 19.0 0.082110 female
7 Retail Store Manager 6.0 -0.002756 male
8 Seasonal Sales Associate 53.0 -0.109181 female
9 Seasonal Sales Associate 26.0 0.000000 male
10 other_jobTitles 146.0 0.231569 female
11 other_jobTitles 125.0 0.613314 male
慕萊塢森
TA貢獻(xiàn)1810條經(jīng)驗(yàn) 獲得超4個(gè)贊
pd.melt(df, id_vars=['jobTitle','mean_z_score_female','mean_z_score_male'],
value_vars=['female_count', 'male_count'], var_name="gender", value_name='count').melt(
id_vars=['jobTitle', 'gender','count'],value_vars=['mean_z_score_female','mean_z_score_male'],
value_name='mean_z_score').drop('variable', axis=1)
慕慕森
TA貢獻(xiàn)1856條經(jīng)驗(yàn) 獲得超17個(gè)贊
這是另一種使用方式lreshape：
newdf=pd.lreshape(df,{'count': ['female_count', 'male_count'],'mean_z_score': ['mean_z_score_female', 'mean_z_score_male']})\
? ? ? ? .sort_values('jobTitle')
newdf['genre']=['female','male']*(len(newdf)//2)
輸出：
newdf
? ? ? ? ? ? ? ? ? ? jobTitle? count? mean_z_score? ?genre
0? ? ? ? ? ? ? ? ? Associate? ?65.0? ? ? 0.000000? female
6? ? ? ? ? ? ? ? ? Associate? ?44.0? ? ?-0.047592? ? male
1? ? ? ? ? ? ? ? ? ? ?Intern? ?13.0? ? ? 0.000000? female
7? ? ? ? ? ? ? ? ? ? ?Intern? ?17.0? ? ? 0.000000? ? male
2? ? ? ? ? ? ? ? ?Key Holder? ?51.0? ? ?-0.352018? female
8? ? ? ? ? ? ? ? ?Key Holder? ?32.0? ? ?-0.288726? ? male
3? ? ? ?Retail Store Manager? ?19.0? ? ? 0.082110? female
9? ? ? ?Retail Store Manager? ? 6.0? ? ?-0.002756? ? male
4? ?Seasonal Sales Associate? ?53.0? ? ?-0.109181? female
10? Seasonal Sales Associate? ?26.0? ? ? 0.000000? ? male
5? ? ? ? ? ? other_jobTitles? 146.0? ? ? 0.231569? female
11? ? ? ? ? ?other_jobTitles? 125.0? ? ? 0.613314? ? male
注意：lreshape目前沒有記錄，可能會(huì)被刪除。
添加回答
舉報(bào)
0/150
提交
取消