ramitg254 commented on code in PR #6413:
URL: https://github.com/apache/hive/pull/6413#discussion_r3438282608
##########
ql/src/java/org/apache/hadoop/hive/ql/metadata/Table.java:
##########
@@ -728,8 +750,50 @@ private boolean isField(String col) {
return false;
}
+ private void ensureColumnsIndexed() {
+ if (columnsByName != null) {
+ return;
+ }
+ Map<String, TableColumn> indexedColumns = new HashMap<>();
+ List<FieldSchema> fsList = new ArrayList<>(getColsInternal(false));
+ if (!hasNonNativePartitionSupport() || isView()) {
+ fsList.addAll(getPartitionKeys());
+ }
+ for (int i = 0; i < fsList.size(); i++) {
+ indexedColumns.put(fsList.get(i).getName().toLowerCase(), new
TableColumn(i, fsList.get(i)));
+ }
+ columnsByName = indexedColumns;
+ }
+
+ public Integer getColumnIndexByName(String colName) {
+ ensureColumnsIndexed();
+ TableColumn column = columnsByName.get(colName.toLowerCase());
+ return column != null ? column.index() : null;
+ }
Review Comment:
I don't mind adding exception like:
```
if (column == null) {
throw new SemanticException(ErrorMsg.INVALID_COLUMN.getMsg(" '" +
colName + "'"));
}
```
but adding this will cascade into lot of places where it is getting used as
getPartCols() also uses getFieldSchemaByName so if we throw exception then we
need cascade it to every location where getPartCols() is used.
I think we can avoid it for now as currently we don't use it in any places
where it can be null.
@deniskuzZ WDYT?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
