shangxinli commented on a change in pull request #950: URL: https://github.com/apache/parquet-mr/pull/950#discussion_r830423474
########## File path: parquet-hadoop/src/main/java/org/apache/parquet/hadoop/ParquetFileReader.java ########## @@ -878,11 +880,92 @@ 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) { paths.clear(); - for (ColumnDescriptor col : projection.getColumns()) { + HashSet<Type.ID> ids = new HashSet<>(); + boolean fileSchemaIdUnique = uniqueId(fileMetaData.getSchema(), ids); + ids = new HashSet<>(); + boolean projectionSchemaIdUnique = uniqueId(projection, ids); + MessageType schema = null; + // if ids are unique, use id resolution Review comment: Instead of figuring out by parquet-mr internals, did you consider letting the caller enable it explicitly? By doing that, we can still check the uniqueness, if not unique, we can throw exceptions. -- 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: dev-unsubscr...@parquet.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org