sqoop-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jarek Jarcec Cecho <jar...@apache.org>
Subject Re: SQOOP INCREMENTAL PULL ISSUE (PLEASE SUGGEST.)
Date Sun, 12 Jan 2014 17:38:16 GMT
Hi Yogesh,
I would start by verifying imported data. If there are duplicates than it's suggesting some
miss configuration of Sqoop, otherwise you might have some inconsistencies down the pipeline.

Jarcec

On Sat, Jan 11, 2014 at 11:01:22PM +0530, yogesh kumar wrote:
> Hello All,
> 
> I am working on a use case where I have to run a process on daily basis
> which will do these.
> 
> 1)  Pull every day new data inserted into RDBMS tables to HDFS
> 2)  Having external table in hive (pointing to the location of HDFS
> directry where data is pulled by sqoop)
> 3) Perform some hive queries (joins) and create a final internal table into
> Hive (say.. Hive_Table_Final).
> 
> 
> What I am doing..
> 
> I am migrating a process from RDBMS to HADOOP ( same process is being
> executed in RDBMS procedure and stored in final table . {say..
>  Rdbms_Table_Final} )
> 
> Issue I am facing is.
> 
> Every time I do Incremental import and after processing I find the final
> table in hive having the value multiplied by every time I do incremental
> import (If I do incremental import to bring new data into HDFS , the data
> in final table of hive after processing  i.e "Hive_Table_Final"  showing
> the values of all columns multiplied by the times of I done incremental
> pull), if I do perform incremental import for 4 days ( every day once
> incremental import in a day and did it for  4 days) i got  data multiplied
> 4 in the final table of hive (Hive_Table_Final)  with respect to final
> table in RDBMS (Rdbms_final_table).
> 
> 
> Like..
> 
> 1) 1st time I have pulled the data from RDBMS based on the months (like
> from 2013-12-01 to 2013-01-01) and processed it, got perfect results
> matching the data in final Hive's  table(Hive_Table_Final) and RDBMS
> processed data into (Rdbms_Table_Final)
> 
> 2) I have done incremental import to bring new data from RDBMS to HDFS by
> using this command..
> 
> 
>  sqoop import -libjars
>  --driver com.sybase.jdbc3.jdbc.SybDriver \
>  --query "select * from
>  from EMP where \$CONDITIONS and SAL > 50000 and SAL <= 80000" \
> --check-column Unique_value \
>  --incremental append \
>  --last-value 201401200 \
>  --split-by DEPT \
>  --fields-terminated-by ',' \
>  --target-dir ${TARGET_DIR}/${INC} \
>  --username ${SYBASE_USERNAME} \
>  --password ${SYBASE_PASSWORD} \
> 
> "Note -- The field Unique_value is very unique for every time, its
> like primary key "
> 
> 
> 
> As now I have just pulled the new records to my HDFS which were into  RDBMS
> tables..
> 
> Now I got major data mis-match issue,  after the
> processing..(Hive_Table_final)
> 
> My Major issue is with sqoop incremental import, as many times I do
> Incremental import I find the  data into my final table gets multiplied by
> the times I have done incremental import..
> 
> 
> Please suggest, whats wrong I am doing, Whats I am missing..
> pls help me out..
> 
> 
> Thanks & Regards
> Yogesh Kumar

Mime
View raw message