rdblue commented on code in PR #950:
URL: https://github.com/apache/parquet-mr/pull/950#discussion_r842094667
##########
parquet-hadoop/src/main/java/org/apache/parquet/hadoop/ParquetFileReader.java:
##########
@@ -878,11 +880,97 @@ public String getFile() {
return blocks;
}
- public void setRequestedSchema(MessageType projection) {
+ private boolean uniqueId(GroupType schema, HashSet<Type.ID> ids) {
+ boolean unique = true;
+ List<Type> fields = schema.getFields();
+ for (Type field : fields) {
+ if (field instanceof PrimitiveType) {
+ Type.ID id = field.getId();
+ if (id != null) {
+ if (ids.contains(id)) {
+ return false;
+ }
+ ids.add(id);
+ }
+ }
+
+ if (field instanceof GroupType) {
+ Type.ID id = field.getId();
+ if (id != null) {
+ if (ids.contains(id)) {
+ return false;
+ }
+ ids.add(id);
+ }
+ if (unique) unique = uniqueId(field.asGroupType(), ids);
+ }
+ }
+ return unique;
+ }
+
+ public MessageType setRequestedSchema(MessageType projection, boolean
useColumnId) {
paths.clear();
- for (ColumnDescriptor col : projection.getColumns()) {
+ MessageType schema = null;
+ if (useColumnId) {
+ HashSet<Type.ID> ids = new HashSet<>();
+ boolean fileSchemaIdUnique = uniqueId(fileMetaData.getSchema(), ids);
+ if (!fileSchemaIdUnique) {
+ throw new RuntimeException("can't use column id resolution because
there are duplicate column ids.");
Review Comment:
Nit: error messages should use sentence case and begin with a capital
letter. I usually like to also show context, like the IDs that are not unique
and the column names within the file.
--
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]
