[
https://issues.apache.org/jira/browse/HIVE-10657?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Alexander Pivovarov updated HIVE-10657:
---------------------------------------
Description:
Current MD5 UDF implementation uses Apache Commons DigestUtils.md5Hex method
to get md5 hex.
DigestUtils does not provide md5Hex method with signature (byte[], start,
length). This is why copyBytes method was added to UDFMd5 to get bytes[] from
BytesWritable.
To avoid copying bytes from BytesWritable to new byte array we can use java
MessageDigest API directly.
MessageDigest has method update(byte[], start, length)
was:
Current implementation uses Apache Commons DigestUtils.md5Hex method to get
md5 hex.
DigestUtils does not provide md5Hex method with signature (byte[], start,
length). This is why copyBytes method was added to get bytes[] from
BytesWritable.
To avoid copying bytes from BytesWritable to new byte array we can use java
MessageDigest API directly.
MessageDigest has method update(byte[], start, length)
> Remove copyBytes operation from MD5 UDF
> ---------------------------------------
>
> Key: HIVE-10657
> URL: https://issues.apache.org/jira/browse/HIVE-10657
> Project: Hive
> Issue Type: Improvement
> Components: UDF
> Reporter: Alexander Pivovarov
> Assignee: Alexander Pivovarov
> Priority: Minor
>
> Current MD5 UDF implementation uses Apache Commons DigestUtils.md5Hex method
> to get md5 hex.
> DigestUtils does not provide md5Hex method with signature (byte[], start,
> length). This is why copyBytes method was added to UDFMd5 to get bytes[] from
> BytesWritable.
> To avoid copying bytes from BytesWritable to new byte array we can use java
> MessageDigest API directly.
> MessageDigest has method update(byte[], start, length)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
