[
https://issues.apache.org/jira/browse/TRAFODION-1457?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14700393#comment-14700393
]
ASF GitHub Bot commented on TRAFODION-1457:
-------------------------------------------
Github user DaveBirdsall commented on a diff in the pull request:
https://github.com/apache/incubator-trafodion/pull/60#discussion_r37247350
--- Diff: core/sql/sqlcomp/CmpSeabaseDDLcommon.cpp ---
@@ -6137,53 +6138,82 @@ short
CmpSeabaseDDL::validateDivisionByExprForDDL(ItemExpr *divExpr)
}
short CmpSeabaseDDL::createEncodedKeysBuffer(char** &encodedKeysBuffer,
+ int &numSplits,
desc_struct * colDescs,
desc_struct * keyDescs,
- Lng32 numSplits, Lng32
numKeys,
- Lng32 keyLength, NABoolean
isIndex)
+ int numSaltPartitions,
+ Lng32 numSaltSplits,
+ NAString *splitByClause,
+ Lng32 numKeys,
+ Lng32 keyLength,
+ NABoolean isIndex)
{
encodedKeysBuffer = NULL;
-
- if (numSplits <= 0)
+ numSplits = 0;
+
+ if (numSaltSplits <= 0)
return 0;
- NAString ** inArray = createInArrayForLowOrHighKeys(colDescs,
- keyDescs,
- numKeys,
- FALSE,
- isIndex,
- STMTHEAP );
-
- char splitNumCharStr[5];
- NAString splitNumString;
-
- /* HBase creates 1 more split than
- the number of rows in the split array. In the example below we have
a
- salt column and an integer column as the key. KeyLength is 4 + 4 =
8.
- encodedKeysBuffer will have 4 elements, each of length 8. When this
- buffer is given to HBase through the Java API we get a table with 5
- splits and begin/end keys as shown below
-
- Start Key End
Key
- \x00\x00\x00\x01\x00\x00\x00\x00
- \x00\x00\x00\x01\x00\x00\x00\x00
\x00\x00\x00\x02\x00\x00\x00\x00
- \x00\x00\x00\x02\x00\x00\x00\x00
\x00\x00\x00\x03\x00\x00\x00\x00
- \x00\x00\x00\x03\x00\x00\x00\x00
\x00\x00\x00\x04\x00\x00\x00\x00
- \x00\x00\x00\x04\x00\x00\x00\x00
- */
- encodedKeysBuffer = new (STMTHEAP) char*[numSplits];
- for(int i =0; i < numSplits; i++)
- encodedKeysBuffer[i] = new (STMTHEAP) char[keyLength];
-
- inArray[0] = &splitNumString;
- short retVal = 0;
-
- for(Int32 i =0; i < numSplits; i++) {
- sprintf(splitNumCharStr, "%d", i+1);
- splitNumString = splitNumCharStr;
+ // make a list of NAStrings with the default values for start keys
+ NAString ** defaultSplits = createInArrayForLowOrHighKeys(
+ colDescs,
+ keyDescs,
+ numKeys,
+ FALSE,
+ isIndex,
+ STMTHEAP );
+ NAString **splitValuesAsText = new(STMTHEAP) NAString *[numKeys];
+ ElemDDLPartitionList * pPartitionList = NULL;
+ ElemDDLPartitionRange *pPartitionRange = NULL;
+
+ numSplits = numSaltSplits;
+
+ // for salt splits, only the first key column value
+ // is variable, the rest use the default values
+ for (int k=1; k<numKeys; k++)
+ splitValuesAsText[k] = defaultSplits[k];
+
--- End diff --
Indentation is inconsistent (if you happen to fix something else perhaps
you could fix this too)
> Fix a collection of issues related to decoding HBase region keys and updating
> the TEXT metadata table
> -----------------------------------------------------------------------------------------------------
>
> Key: TRAFODION-1457
> URL: https://issues.apache.org/jira/browse/TRAFODION-1457
> Project: Apache Trafodion
> Issue Type: Bug
> Environment: any
> Reporter: Hans Zeller
> Assignee: Hans Zeller
>
> The code I wrote a while ago for handling incomplete HBase region keys needs
> some more improvement, which caused me to clean up a few related areas. One
> of them is generation of min/max values for key building. Also, I would like
> to change the methods we use to read and update the "_MD_".TEXT table.
> Changing quotes to double quotes should happen inside this method, not last
> because it should only happen on the individual chunks to be inserted.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
