[GEODE-9174] The result of a gfsh query containing a UUID may not be displayed properly

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.15.0
• Component/s: gfsh, querying
• Labels: pull-request-available
• Assignee: Mario Kevo
• Reporter: Barrett Oglesby
• Created: 19/Apr/21 21:04
• Updated: 22/Jun/22 20:46
• Resolved: 28/Apr/21 06:17

Description

For example, if the key is a UUID, then a query like this won't show the results even though there is one:

gfsh>query --query="select key from /data.entries where value.id = '55e907b6-a1fe-42ea-90a2-6a5698e9b27c'"
Result : true
Limit  : 100
Rows   : 1

But a query like this will:

gfsh>query --query="select key,value from /data.entries where value.id = '55e907b6-a1fe-42ea-90a2-6a5698e9b27c'"
Result : true
Limit  : 100
Rows   : 1

                 key                   | value
-------------------------------------- | ---------------------------------------------------------------------------------------
"55e907b6-a1fe-42ea-90a2-6a5698e9b27c" | {"id":"55e907b6-a1fe-42ea-90a2-6a5698e9b27c","cusip":"AAPL","shares":22,"price":352.32}

Thats because of the way DataCommandResult.resolveObjectToColumns works.

private void resolveObjectToColumns(Map<String, String> columnData, Object value) {
  if (value instanceof PdxInstance) {
    resolvePdxToColumns(columnData, (PdxInstance) value);
  } else if (value instanceof Struct) {
    resolveStructToColumns(columnData, (StructImpl) value);
  } else {
    ObjectMapper mapper = new ObjectMapper();
    JsonNode node = mapper.valueToTree(value);
    node.fieldNames().forEachRemaining(field -> {
      ...
      columnData.put(field, mapper.writeValueAsString(node.get(field)));
    });
  }
}

The value in the first query is a UUID so the last else clause is invoked. In this case, a JsonNode is used to determine the columns. ObjectMapper.valueToTree converts a UUID to a TextNode. TextNodes have no fieldNames, and JsonNode.fieldNames returns an EmptyIterator by default:

public Iterator<String> fieldNames() {
  return ClassUtil.emptyIterator();
}

So, resolveObjectToColumns doesn't fill in columnData, which causes the DataCommandResult.buildTable in the locator to not add any rows to the table.

The value in the second query is a Struct so the second else clause is invoked. The resolveStructToColumns method does:

private void resolveStructToColumns(Map<String, String> columnData, StructImpl struct) {
  for (String field : struct.getFieldNames()) {
    columnData.put(field, valueToJson(struct.get(field)));
  }
}

I'm not sure if there is a way to make ObjectMapper.valueToTree handle UUIDs differently, but they can easily be special-cased like PdxInstances and Structs:

} else if (value instanceof UUID) {
  columnData.put("uuid", valueToJson(value));

I'm not sure if this is the best solution, but it works. With this clause added, the query does:

gfsh>query --query="select key from /data.entries where value.id = '55e907b6-a1fe-42ea-90a2-6a5698e9b27c'"
Result : true
Limit  : 100
Rows   : 1

uuid
--------------------------------------
"55e907b6-a1fe-42ea-90a2-6a5698e9b27c"

Issue Links

• links to: GitHub Pull Request #6365