Replication issue in rowcount

[vuppalavenu]

Hi, I am facing one typical issue.  When I am loading from SQL Server to Filesystem, it transfers all the records, like 129 rows.  When I use the same file on SQL Server, it transfers only 126 rows.  Please help.

[Steve_Nguyen]

@vuppalavenu 

when you check the file , does it have header and footer , etc?

I use the same file on SQL Server,,,, what do you mean ?

129 rows is easy to check , what record are missing ?  

[vuppalavenu]

Thanks, Steve  Here we divided it into 2  steps instead of direct replication.  First step SQL Server to FileChannel then FileChannel to SqlServer another server.  FileChanel file is not readable.  Please suggest.

[Steve_Nguyen]

@vuppalavenu 

1. so from SQL to Filechannel and Filechannel to SQL,, you get 126 row ,,, correct ?

2. if you try SQL to Filechannel and Filechannel to Mysql , Oracle, or other ,, what do you get ?

where are you seeing the 129 rows?

[vuppalavenu]

1. From SQL to FileChannel  Qlick Replicate shows 129 rows transferred.  Source we have 129 rows.

2. From FileChannel to SQL server Qlick Replicate shows 126 rows transferred.  User confirmed they are seeing only 126 rows .

[Heinvandenheuvel · Accepted solution]

As @Steve_Nguyen indicates - with just 100 rows or so you could even manually check which PK's are not present on target and see what is special with those rows (Like a NULL or a strange date). I'd script it or extract all PK's source and Target and run a (win)DIFF.

How is the error handling set up? What happened to the missing rows? Is there a mention in the log? Are there entries for them in the apply_exceptions table?

Hein.

[vuppalavenu]

Thanks, Steve You are right. Due to the primary key issue, records do not flow, and the data flowed correctly after removing the primary key.