官方文档 :
DataComPy — datacompy 0.8.2 documentationhttps://capitalone.github.io/datacompy/index.html要求:
这2个DataFrame的列是完全一样,列数与列名。否则会报错:
KeyError: 'xxxxx_match'
如是列名不一样可以先重命名列名再比较。然后根据列名取出比较的结果,返回的都是DataFrame
df1.columns=[1,2,3,4,5]
df2.columns=[1,2,3,4,5]
dd = datacompy.Compare(df1, df2, join_columns="1") # '1' 是列名
print(dd.report()) #打印所有比较结果
print('---- 2----')
diff_per = dd.sample_mismatch('2') #
print(diff_per)
print('---- 3----')
diff_per1 = dd.sample_mismatch('3')
print(diff_per1) # 取出列3不相同的数据,返回 dataframe
print('---- 4-----')
diff_per2 = dd.sample_mismatch('4')
print(diff_per) # 取出不相同的某一列数据,返回 dataframe
print('---- df1 新增-----')
print(dd.df1_unq_rows)
print('---- df2 新增-----')
print(dd.df2_unq_rows)
dd.report() 结果:
DataComPy Comparison
--------------------DataFrame Summary
-----------------DataFrame Columns Rows
0 df1 5 53
1 df2 5 50Column Summary
--------------Number of columns in common: 5
Number of columns in df1 but not in df2: 0
Number of columns in df2 but not in df1: 0Row Summary
-----------Matched on: 1
Any duplicates on match values: No
Absolute Tolerance: 0
Relative Tolerance: 0
Number of rows in common: 43
Number of rows in df1 but not in df2: 10
Number of rows in df2 but not in df1: 7Number of rows with some compared columns unequal: 8
Number of rows with all compared columns equal: 35Column Comparison
-----------------Number of columns compared with some values unequal: 4
Number of columns compared with all values equal: 1
Total number of values which compare unequal: 13Columns with Unequal Values or Types
------------------------------------Column df1 dtype df2 dtype # Unequal Max Diff # Null Diff
0 2 object object 3 0.0 1
1 3 float64 float64 4 25.0 1
2 4 object object 3 0.0 1
3 5 object object 3 0.0 1Sample Rows with Unequal Values
-------------------------------1 2 (df1) 2 (df2)
...
Sample Rows Only in df1 (First 10 Columns)
------------------------------------------1 2 3 4 5
...
Sample Rows Only in df2 (First 10 Columns)
------------------------------------------1 2 3 4 5
datacompy.Compare()参数:
# Compare 参数:
# df1: 数据框1
# df2: 数据框2
# join_columns: 指定索引的列名,默认“None”,可以传入数组,比如:['ID', 'Name']
# on_index: 是否要开启索引,开启之后不需要指定 join_columns,默认“False”
# abs_tol: 绝对公差,默认“0”
# rel_tal: 相对公差,默认“0”
# df1_name: 报告中数据框1的名字,默认“df1”
# df2_name: 报告中数据框2的名字,默认“df2”
# ignore_spaces: 是否忽略空格,默认“False”
# ignore_case: 是否忽略大小写,默认“False”
问题:生成df2不在df1的数据,row的索引值与原始表对不上。
解决:如想获取到真正的原表的df2不在df1的行索引。可以反转df1,df2的比较顺序再做一次比较 ,
dd2 = datacompy.Compare(df2, df1, join_columns="1") # '1' 是列名
然后再打印:
print('---- df2 新增的行-----') print(dd2.df1_unq_rows) # 依然是用:df1_unq_rows, 按入参顺序。比较结果第一个df2会在datacompy里面的值为df1