如何在行级别比较两个海量火花数据帧并打印差异

Question

I have two very large spark dataframes. I want to compare it on row level and print only the differences

eg:

df1= firstname:abc lastname:xyz company:123

---

df2= firstname:abc lastname:xyz company:456

---

expected output- diff= company(df1):123 company(df2):456

Answer 1

As of me,there is no optimal solution for the problem you have described. Because, a difference between dataframes can be found only when you have a column/reference on which both the dataframes could be joined.

With that note, one approach will be to use subtract function find the difference which helps to some extent.

>>> df_1.show()
+-----+-----+-----+
|fname|lname|cmpny|
+-----+-----+-----+
|  abc|  xyz|  123|
+-----+-----+-----+

>>> df_2.show()
+-----+-----+-----+
|fname|lname|cmpny|
+-----+-----+-----+
|  abc|  xyz|  456|
+-----+-----+-----+

>>> df_1.select('*').subtract(df_2.select('*')).show()
+-----+-----+-----+
|fname|lname|cmpny|
+-----+-----+-----+
|  abc|  xyz|  123|
+-----+-----+-----+

>>> df_2.select('*').subtract(df_1.select('*')).show()
+-----+-----+-----+
|fname|lname|cmpny|
+-----+-----+-----+
|  abc|  xyz|  456|
+-----+-----+-----+

Answer 2

I think you are looking for except

df1.except(df2)

will return rows in df1 not in df2.