sqoop-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Abraham Elmahrek <...@cloudera.com>
Subject Re: Binary data transfer using Sqoop
Date Thu, 16 Jul 2015 05:02:34 GMT
http://sqoop.apache.org/docs/1.4.6/SqoopUserGuide.html#_large_objects

use --inline-lob-limit 0.

-Abe

On Wed, Jul 15, 2015 at 9:22 PM, Suraj Nayak <snayakm@gmail.com> wrote:

> Hi All,
>
> I tried --map-column-java com.cloudera.sqoop.lib.BlobRef by using --query
> "select CAST(email as BLOB) from table_name where \$CONDITIONS", and
> --map-column-java org.apache.hadoop.io.BytesWritable. Still the same
> issue.
>
>
> Sqoop Output: 20 16 3a 34 fa 40 f6 fc 24 3d 5f b9 c6 63 b5 5f 54 77
> Required output : [image: Inline image 1]
>
>
>
>
>
> On Wed, Jul 15, 2015 at 9:04 PM, Suraj Nayak <snayakm@gmail.com> wrote:
>
>> One correction in Sqoop version. Its 1.4.4 not 1.4.5.
>>
>> On Wed, Jul 15, 2015 at 8:56 PM, Suraj Nayak <snayakm@gmail.com> wrote:
>>
>>> Is it by setting --map-column-java email=com.cloudera.sqoop.lib.BlobRef?
>>>
>>> On Wed, Jul 15, 2015 at 8:11 PM, Suraj Nayak <snayakm@gmail.com> wrote:
>>>
>>>> Hi Abe,
>>>>
>>>> I think the 1st option suits my use case best. But, what is the
>>>> method/way to tell Sqoop to use Large Object Loading?
>>>>
>>>> Thanks!
>>>>
>>>> On Wed, Jul 15, 2015 at 7:50 PM, Abraham Elmahrek <abe@cloudera.com>
>>>> wrote:
>>>>
>>>>> Hey man,
>>>>>
>>>>> I think Sqoop does this intentionally:
>>>>> https://github.com/apache/sqoop/blob/trunk/src/java/org/apache/sqoop/mapreduce/RawKeyTextOutputFormat.java#L65
>>>>> .
>>>>>
>>>>> I see two things that may get you beyond this point:
>>>>>
>>>>>    1. You can try type casting to a BLOB and tell Sqoop to use Large
>>>>>    Object Loading.
>>>>>    2. Try the --as-avrodatafile option and write as Avro.
>>>>>
>>>>> -Abe
>>>>>
>>>>> On Wed, Jul 15, 2015 at 7:23 PM, Suraj Nayak <snayakm@gmail.com>
>>>>> wrote:
>>>>>
>>>>>> Hi Abe,
>>>>>>
>>>>>> Thanks for highlighting missing required info quickly. Below are
the
>>>>>> details:
>>>>>>
>>>>>>    - *Version:* Sqoop 1.4.5
>>>>>>    - *Sqoop Command: *sqoop import --connect
>>>>>>    jdbc:teradata://aa.bb.cc.internal/DATABASE=someDB --username sqoop_usr
>>>>>>    --password sqoop_usr --table ENCRYPTED_TBL --fields-terminated-by
\\001 -m
>>>>>>    1 --target-dir /tmp/ENC_TBL --connection-manager
>>>>>>    "org.apache.sqoop.manager.GenericJdbcManager" --driver
>>>>>>    com.teradata.jdbc.TeraDriver
>>>>>>    - *Table structure:* id:varchar, count:int, first_name:binary,
>>>>>>    email:binary, column5:varchar.
>>>>>>
>>>>>> Binary is used as the data is encrypted.
>>>>>>
>>>>>> Thanks!
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Wed, Jul 15, 2015 at 6:44 PM, Abraham Elmahrek <abe@cloudera.com>
>>>>>> wrote:
>>>>>>
>>>>>>> Hey man,
>>>>>>>
>>>>>>> Need some details to help:
>>>>>>>
>>>>>>>    - What version of Sqoop?
>>>>>>>    - Sqoop command?
>>>>>>>    - Database table structure (preferably a describe on the database)
>>>>>>>
>>>>>>> -Abe
>>>>>>>
>>>>>>> On Wed, Jul 15, 2015 at 6:42 PM, Suraj Nayak <snayakm@gmail.com>
>>>>>>> wrote:
>>>>>>>
>>>>>>> > Hi Sqoop Users and Developers,
>>>>>>> >
>>>>>>> > How can i import a Binary data column in a table into HDFS
without
>>>>>>> > converting it into String.
>>>>>>> >
>>>>>>> > I have encrypted data in RDBMS, I need to import this column
as is
>>>>>>> without
>>>>>>> > converting it into string. As of now, Sqoop is typecasting
the
>>>>>>> data into
>>>>>>> > String/text and decryption is failing in Hadoop.
>>>>>>> >
>>>>>>> > Can someone provide pointers to solve this? Any workaround?
>>>>>>> >
>>>>>>> > --
>>>>>>> > Thanks
>>>>>>> > Suraj Nayak M
>>>>>>> >
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Thanks
>>>>>> Suraj Nayak M
>>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> Thanks
>>>> Suraj Nayak M
>>>>
>>>
>>>
>>>
>>> --
>>> Thanks
>>> Suraj Nayak M
>>>
>>
>>
>>
>> --
>> Thanks
>> Suraj Nayak M
>>
>
>
>
> --
> Thanks
> Suraj Nayak M
>

Mime
  • Unnamed multipart/related (inline, None, 0 bytes)
View raw message