Good Day!
We are required to convert the CSV data to be stored in parquet format. While converting using the below NiFi flow, we see .member0 & .member1 getting added to the output.
Input Data :
no,firstname,lastname
1,1234,user
1,test,user
1,test,user
1,test,user
1,test,user
2,babu,user2
c,krish,mohan
4,krishna1234,mohna
NiFi Flow :
Output Data :
no:
.member0 = 1
firstname:
.member0 = 1234
lastname = user
no:
.member0 = 1
firstname:
.member1 = test
lastname = user
no:
.member0 = 1
firstname:
.member1 = test
lastname = user
no:
.member0 = 1
firstname:
.member1 = test
lastname = user
no:
.member0 = 1
firstname:
.member1 = test
lastname = user
no:
.member0 = 2
firstname:
.member1 = babu
lastname = user2
no:
.member1 = c
firstname:
.member1 = krish
lastname = mohan
no:
.member0 = 4
firstname:
.member1 = krishna1234
lastname = mohna
We believe internally somehow the integer values are assigned to member0 and string values are assigned to member1 if the same column has both string and integer values. Also, we believe this is decided based on the values of 1st row. Please help us with your expertise if we can avoid this member0 and member1 by explicitly mentioning to NiFi to treat all data as a string. Thanks in advance!
