Github user zellerh commented on a diff in the pull request:
https://github.com/apache/incubator-trafodion/pull/496#discussion_r64677400
--- Diff: core/sql/optimizer/NATable.cpp ---
@@ -3553,6 +3553,48 @@ NAType* getSQColTypeForHive(const char* hiveType,
NAMemory* heap)
if ( !strcmp(hiveType, "timestamp"))
return new (heap) SQLTimestamp(TRUE /* allow NULL */ , 6, heap);
+ if ( !strcmp(hiveType, "date"))
+ return new (heap) SQLDate(TRUE /* allow NULL */ , heap);
+
+ if ( !strncmp(hiveType, "varchar", 7) )
+ {
+ char maxLen[32];
+ memset(maxLen, 0, 32);
+ int i=0,j=0;
+ int copyit = 0;
+
+ //get length
+ for(i = 0; i < strlen(hiveType) ; i++)
+ {
+ if(hiveType[i] == '(') //start
+ {
+ copyit=1;
+ continue;
+ }
+ else if(hiveType[i] == ')') //stop
+ break;
+ if(copyit > 0)
+ {
+ maxLen[j] = hiveType[i];
+ j++;
+ }
+ }
+ Int32 len = atoi(maxLen);
+
+ if(len == 0) return NULL; //cannot parse correctly
+
+ NAString hiveCharset =
+ ActiveSchemaDB()->getDefaults().getValue(HIVE_DEFAULT_CHARSET);
+
+ return new (heap) SQLVarChar(CharLenInfo((hiveCharset ==
CharInfo::UTF8 ? 0 : len),len),
--- End diff --
One solution is to do this mapping:
Hive STRING ==> Trafodion VARCHAR(n BYTES), with n determined by
HIVE_MAX_STRING_LENGTH
Hive VARCHAR(n) ==> Trafodion VARCHAR(n)
This treats STRING and VARCHAR differently, the max size of STRING is in
bytes, the max size of VARCHAR is in characters. This is not ideal, users will
probably expect string + CQD HIVE_MAX_STRING_LENGTH 'n' to work similar to
VARCHAR(n)?
There is no good way to achieve this, though, at least not without wasting
space and making incompatible changes.
I wonder what other people think.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
