sqoop-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Gwen Shapira <gshap...@cloudera.com>
Subject Re: Import NULL value HBase
Date Tue, 29 Jul 2014 15:30:46 GMT
Let us know how it goes. If there are performance issues, we may be
able to fix them :)

On Tue, Jul 29, 2014 at 7:34 AM, Alessandra Tiengo <atiengo@daisylabs.it> wrote:
> Thanks for your reply!
> I read that using Sqoop to import from MySQL directly to HBase is not very performant,
but I will try it.
>
> Alessandra
>
> -----Original Message-----
> From: Gwen Shapira [mailto:gshapira@cloudera.com]
> Sent: venerdì 25 luglio 2014 19:24
> To: user@sqoop.apache.org
> Subject: Re: Import NULL value HBase
>
> Any reason you are not using Sqoop to import from MySQL directly to HBase? This seems
like the easiest solution.
>
> Other possible solutions:
> - Use the empty string as null value
> - Modify ImportTSV so it will skip the value "null" in a text file.
>
> Gwen
>
> On Fri, Jul 25, 2014 at 8:05 AM, Alessandra Tiengo <atiengo@daisylabs.it> wrote:
>> Hi,
>>
>> I imported a table in HDFS from a SQL database using SQOOP. Null
>> values have been replaced by “null” string, so a row is (for
>> instance): rowkey, value1, value2, null, null
>>
>>
>>
>> After that I imported the data from HDFS to a HBase table using first
>> org.apache.hadoop.hbase.mapreduce.ImportTsv and then
>> org.apache.hadoop.hbase.mapreduce.LoadIncrementalHFiles.
>>
>> How can I manage these “null” values so that they are not imported in
>> the HBase table?
>>
>> Currently they are present as “null” string.
>>
>>
>>
>> I also tried to use --null-string ‘\\N’ --null-non-string ‘\\N’, but
>> HBase stored \N.
>>
>>
>>
>> Thanks,
>>
>>
>>
>> Alessandra
>>
>>
>>

Mime
View raw message