我将三个数据框合并在一起，然后从中删除重复的数据框。但是，当我从最后三列中删除重复项时，我在要删除的数据框的顶部获得了NaN值，但似乎找不到解决方法。
到目前为止，这是我的代码：
bDF=pd.read_csv(bRaw)
pDF=pd.read_csv(pRaw)
mDF=pd.read_csv(mRaw)
del bRaw,pRaw,mRaw
#Merge Together Datarames on the Value Role Name
dfs=[bDF,pDF,mDF]
df_merged = reduce(lambda left,right: pd.merge(left,right,on=['R1'],
how='outer'), dfs)
del bDF,pDF,mDF,dfs
#Rearrange Columns
cols=df_merged.columns.tolist()
cols=cols[0:1]+cols[-3:]+cols[1:5]
df_merged=df_merged[cols]
合并后的输出：
+------+-----+------+----+--------+--------+--------+--------+
| R | C | D | JC | R | PM | Nme | Vle |
+------+-----+------+----+--------+--------+--------+--------+
| JMAC | 305 | 3302 | I6 | Cofow | Value1 | Value1 | Value1 |
| JMAC | 305 | 3915 | R6 | Cofow | Value1 | Value1 | Value1 |
| JMAC | 301 | 3302 | I6 | Cofow | Value1 | Value1 | Value1 |
| JMAC | 301 | 3915 | R6 | Cofow | Value1 | Value1 | Value1 |
| JMAC | 305 | 3302 | I6 | Cofow | Value2 | Value2 | Value2 |
| JMAC | 305 | 3915 | R6 | Cofow | Value2 | Value2 | Value2 |
| JMAC | 301 | 3302 | I6 | Cofow | Value2 | Value2 | Value2 |
| JMAC | 301 | 3915 | R6 | Cofow | Value2 | Value2 | Value2 |
| JMAC | 305 | 3302 | I6 | Cofow | Value3 | Value3 | Value3 |
| JMAC | 305 | 3915 | R6 | Cofow | Value3 | Value3 | Value3 |
| JMAC | 301 | 3302 | I6 | Cofow | Value3 | Value3 | Value3 |
| JMAC | 301 | 3915 | R6 | Cofow | Value3 | Value3 | Value3 |
| JMAC | 305 | 3302 | I6 | Cofow | Value4 | Value4 | Value4 |
| JMAC | 305 | 3915 | R6 | Cofow | Value4 | Value4 | Value4 |
| JMAC | 301 | 3302 | I6 | Cofow | Value4 | Value4 | Value4 |
| JMAC | 301 | 3915 | R6 | Cofow | Value4 | Value4 | Value4 |
| JMAP | 301 | 3315 | I6 | Cofowd | Value6 | Value6 | Value6 |
| JMAP | 301 | 3916 | R6 | Cofowd | Value6 | Value6 | Value6 |
| JMAP | 305 | 3314 | I6 | Cofowd | Value6 | Value6 | Value6 |
| JMAP | 305 | 3315 | R6 | Cofowd | Value6 | Value6 | Value6 |
| JMAP | 305 | 3916 | R6 | Cofowd | Value6 | Value6 | Value6 |
| JMAP | 301 | 3315 | I6 | Cofowd | Value7 | Value7 | Value7 |
| JMAP | 301 | 3916 | R6 | Cofowd | Value7 | Value7 | Value7 |
| JMAP | 305 | 3314 | I6 | Cofowd | Value7 | Value7 | Value7 |
| JMAP | 305 | 3315 | R6 | Cofowd | Value7 | Value7 | Value7 |
| JMAP | 305 | 3916 | R6 | Cofowd | Value7 | Value7 | Value7 |
| JMAP | 301 | 3315 | I6 | Cofowd | Value8 | Value8 | Value8 |
| JMAP | 301 | 3916 | R6 | Cofowd | Value8 | Value8 | Value8 |
| JMAP | 305 | 3314 | I6 | Cofowd | Value8 | Value8 | Value8 |
| JMAP | 305 | 3315 | R6 | Cofowd | Value8 | Value8 | Value8 |
| JMAP | 305 | 3916 | R6 | Cofowd | Value8 | Value8 | Value8 |
| JMAP | 301 | 3315 | I6 | Cofowd | Value9 | Value9 | Value9 |
| JMAP | 301 | 3916 | R6 | Cofowd | Value9 | Value9 | Value9 |
| JMAP | 305 | 3314 | I6 | Cofowd | Value9 | Value9 | Value9 |
| JMAP | 305 | 3315 | R6 | Cofowd | Value9 | Value9 | Value9 |
| JMAP | 305 | 3916 | R6 | Cofowd | Value9 | Value9 | Value9 |
+------+-----+------+----+--------+--------+--------+--------+
然后我从前4列，后三列，最后是中间列中删除重复项：
#Remove Duplicate Values
df_merged[cols[0:-3]]=df_merged[cols[0:-3]].mask(df_merged[cols[:-3]].duplicated())
df_merged[cols[-3:]]=df_merged[cols[-3:]].mask(df_merged[cols[-3:]].duplicated())
df_merged[cols[4:5]]=df_merged[cols[4:5]].mask(df_merged[cols[4:5]].duplicated())
df_merged=df_merged.dropna(how='all')
我的输出接近最终表单的要求：
+------+-----+------+----+-------+---------+---------+---------+
| R | C | D | JC | R | PM | Nme | Vle |
+------+-----+------+----+-------+---------+---------+---------+
| JMAC | 305 | 3302 | I6 | Cofow | Value1 | Value1 | Value1 |
| JMAC | 305 | 3915 | R6 | | NaN | NaN | NaN |
| JMAC | 301 | 3302 | I6 | | NaN | NaN | NaN |
| JMAC | 301 | 3915 | R6 | | NaN | NaN | NaN |
| | | | | | Value2 | Value2 | Value2 |
| | | | | | Value3 | Value3 | Value3 |
| | | | | | Value4 | Value4 | Value4 |
| | | | | | Value6 | Value6 | Value6 |
| | | | | | Value7 | Value7 | Value7 |
| JMAP | 301 | 3315 | I6 | Cofow | Value8 | Value8 | Value8 |
| JMAP | 301 | 3916 | R6 | | NaN | NaN | NaN |
| JMAP | 305 | 3314 | I6 | | NaN | NaN | NaN |
| JMAP | 305 | 3315 | R6 | | NaN | NaN | NaN |
| JMAP | 305 | 3916 | R6 | | NaN | NaN | NaN |
| | | | | | Value9 | Value9 | Value9 |
| | | | | | Value10 | Value10 | Value10 |
| | | | | | Value11 | Value11 | Value11 |
| | | | | | Value12 | Value12 | Value12 |
| | | | | | Value13 | Value13 | Value13 |
+------+-----+------+----+-------+---------+---------+---------+
我的问题是我想摆脱NaN值并向上移动值。所以我希望最终结果看起来像这样：
+------+-----+------+----+-------+---------+---------+---------+
| R | C | D | JC | R | PM | Nme | Vle |
+------+-----+------+----+-------+---------+---------+---------+
| JMAC | 305 | 3302 | I6 | Cofow | Value1 | Value1 | Value1 |
| JMAC | 305 | 3915 | R6 | | Value2 | Value2 | Value2 |
| JMAC | 301 | 3302 | I6 | | Value3 | Value3 | Value3 |
| JMAC | 301 | 3915 | R6 | | Value4 | Value4 | Value4 |
| | | | | | Value6 | Value6 | Value6 |
| | | | | | Value7 | Value7 | Value7 |
| JMAP | 301 | 3315 | I6 | Cofow | Value8 | Value8 | Value8 |
| JMAP | 301 | 3916 | R6 | | Value9 | Value9 | Value9 |
| JMAP | 305 | 3314 | I6 | | Value10 | Value10 | Value10 |
| JMAP | 305 | 3315 | R6 | | Value11 | Value11 | Value11 |
| JMAP | 305 | 3916 | R6 | | Value12 | Value12 | Value12 |
| | | | | | Value13 | Value13 | Value13 |
+------+-----+------+----+-------+---------+---------+---------+
我曾尝试将列分为两个不同的数据帧，以除去NA，然后将它们组合在一起，但是由于索引，我的数据被丢弃了。
df3=pd.concat([df2,df1], axis=1, ignore_index=False)
任何帮助或想法都很棒！
非常感谢，
要旨
答案 0 :(得分：0)
然后我从前4列，后三列，最后是中间列中删除重复项：
假设这些是您要执行的步骤，请尝试drop_duplicates。这是一个示例，它将在一个命令中按您的顺序执行此操作：
df = df.drop_duplicates(
subset=['col1', 'col2', 'col3', 'col4']).drop_duplicates(
subset=['col6', 'col7', 'col8']).drop_duplicates(
subset=['col5'])
您还可以使用keep参数（例如keep='first'与keep='last'）来更改要删除/保留的行。