时间戳

首页 > 解决方案 > 比较多个 Pandas 数据框的所有列名

python - 比较多个 Pandas 数据框的所有列名

问题描述

我正在寻找一种更简单的方法来比较具有与下面代码类似的输出的多个列名，或者寻找一种使此函数更简洁的方法。每次我尝试使用 *args 或 **Kwargs 来简化函数时，都会收到错误“IndexError: list index out of range”或“函数需要 0 个位置参数但给出了 1”。任何意见是极大的赞赏。

'''

application_df = pd.read_csv("application_train.csv")
bureau_df = pd.read_csv("bureau.csv")
bureau_balance_df = pd.read_csv("bureau_balance.csv")
previous_application_df = pd.read_csv("previous_application.csv")
POS_CASH_balance_df = pd.read_csv("POS_CASH_balance.csv")
installments_payments_df = pd.read_csv("installments_payments.csv")
credit_card_balance_df = pd.read_csv("credit_card_balance.csv")
sample__submission_df = pd.read_csv("sample_submission.csv")

def column_compare(df1, df2, df3, df4, df5, df6, df7, df8):
    alist = []
    blist = []
    clist = []
    dlist = []
    elist = []
    flist = []
    glist = []
    hlist = []
    
    name1 =[x for x in globals() if globals()[x] is df1][0]
    for a in df1:
        alist.append(a)
        
    name2 =[x for x in globals() if globals()[x] is df2][0]
    for b in df2:
        blist.append(b)
        
    name3 =[x for x in globals() if globals()[x] is df3][0]
    for c in df3:
        clist.append(c)
        
    name4 =[x for x in globals() if globals()[x] is df4][0]
    for d in df4:
        dlist.append(d)
        
    name5 =[x for x in globals() if globals()[x] is df5][0]
    for e in df5:
        elist.append(e)
        
    name6 =[x for x in globals() if globals()[x] is df6][0]
    for f in df6:
        flist.append(f)
        
    name7 =[x for x in globals() if globals()[x] is df7][0]
    for g in df7:
        glist.append(g)
        
    name8 =[x for x in globals() if globals()[x] is df8][0]
    for h in df8:
        hlist.append(h)
        
    dfs = {name1: alist, name2: blist, name3: clist, name4: dlist, name5: elist, name6: flist, name7: glist, name8: hlist}
    df = pd.DataFrame.from_dict(dfs, orient='index')
    df=df.transpose().replace(np.nan,'')
    return df

pd.set_option("max_rows", None)
column_names = column_compare(application_df, bureau_df, bureau_balance_df, previous_application_df, POS_CASH_balance_df, installments_payments_df, credit_card_balance_df, sample__submission_df)
column_names

'''

标签： pythonpandasdataframecsv

解决方案

不确定这是否是您感兴趣的内容，但可以通过 访问列名df.columns，因此：

import pandas as pd

df1 = pd.DataFrame(columns=list('abc'))
df2 = pd.DataFrame(columns=list('abefg'))
df3 = pd.DataFrame(columns=list('bfkm'))

cols1 = pd.DataFrame(index=df1.columns).assign(df1=1).T
cols2 = pd.DataFrame(index=df2.columns).assign(df2=1).T
cols3 = pd.DataFrame(index=df3.columns).assign(df3=1).T

df_cols_compare = pd.concat([cols1, cols2, cols3]).fillna(0).astype('int')

#      a  b  c  e  f  g  k  m
# df1  1  1  1  0  0  0  0  0
# df2  1  1  0  1  1  1  0  0
# df3  0  1  0  0  1  0  1  1

推荐阅读