Github user merrimanr commented on a diff in the pull request: https://github.com/apache/metron/pull/827#discussion_r149156555 --- Diff: metron-platform/metron-indexing/src/test/java/org/apache/metron/indexing/dao/InMemoryDao.java --- @@ -221,12 +222,24 @@ public void batchUpdate(Map<Document, Optional<String>> updates) throws IOExcept } } - public Map<String, Map<String, FieldType>> getColumnMetadata(List<String> indices) throws IOException { - Map<String, Map<String, FieldType>> columnMetadata = new HashMap<>(); + @Override + public Map<String, FieldType> getColumnMetadata(List<String> indices) throws IOException { + Map<String, FieldType> indexColumnMetadata = new HashMap<>(); for(String index: indices) { - columnMetadata.put(index, new HashMap<>(COLUMN_METADATA.get(index))); + Map<String, FieldType> columnMetadata = COLUMN_METADATA.get(index); + for (Entry entry: columnMetadata.entrySet()) { + String field = (String) entry.getKey(); + FieldType type = (FieldType) entry.getValue(); + if (indexColumnMetadata.containsKey(field)) { + if (!type.equals(indexColumnMetadata.get(field))) { + indexColumnMetadata.remove(field); --- End diff -- I agree with you that it would be confusing but I'm not sure what the correct behavior should be. Should we include the field but just set the type to OTHER? This is how the ElasticsearchDao treats fields it doesn't have type information for but so it might be better to explicitly state this in the column metadata endpoint response.
---