Tim Armstrong has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/14660 )

Change subject: IMPALA-9090 Add name of table being scanned in scan node profile
......................................................................


Patch Set 2:

(2 comments)

http://gerrit.cloudera.org:8080/#/c/14660/2//COMMIT_MSG
Commit Message:

http://gerrit.cloudera.org:8080/#/c/14660/2//COMMIT_MSG@24
PS2, Line 24: "test.test.hbase_table".
This seems buggy - I looked at "struct TTableDescriptor" and it says that 
tableName should be the unqualified name of the table, presumably the HMS table 
name, but I think the HbaseTable class in java is passing in the hbase table 
name incorrectly:

        new TTableDescriptor(tableId, TTableType.HBASE_TABLE, 
getTColumnDescriptors(),
            numClusteringCols_, hbaseTableName_, db_.getName());


It looks like table_name() is meant to be the native name of the Kudu and HBase 
tables. I think a few places in the kudu and hbase backend code are incorrectly 
using name() instead of table_name().

  aggregator.cc:      runtime_profile_, -1, runtime_profile_->name(), 
exec_node_->mem_tracker()));
  data-source-scan-node.cc:    
col.__set_name(tuple_desc_->table_desc()->col_descs()[col_idx].name());
  data-source-scan-node.cc:  
params.__set_table_name(tuple_desc_->table_desc()->name());
  exec-node.cc:  mem_tracker_.reset(new MemTracker(runtime_profile_, -1, 
runtime_profile_->name(),
  hbase-table-writer.cc:  
RETURN_IF_ERROR(ExecEnv::GetInstance()->htable_factory()->GetTable(table_desc_->name(),
  hdfs-avro-scanner.cc:      
scan_node_->hdfs_table()->col_descs()[slot_desc->col_pos()].name();
  hdfs-scan-node-base.cc:          hdfs_table_->database(), 
hdfs_table_->name())));
  hdfs-table-sink.cc:    partition_name_ss << 
table_desc_->col_descs()[j].name() << "=";
  kudu-scanner.cc:           << " Kudu table=" << scan_node_->table_->name();
  kudu-scanner.cc:              scan_node_->table_->name(),
  kudu-scanner.cc:              
scan_node_->table_->schema().Column(slot->col_pos()).name())));
  kudu-scanner.cc:      scan_node_->table_->name());
  kudu-table-sink.cc:          "Table $0 has fewer columns than expected.", 
table_desc_->name()));
  kudu-table-sink.cc:          table_->schema().Column(col_idx).name(), 
type.DebugString(),
  kudu-table-sink.cc:                     << table_->schema().Column(col).name()
  parquet/hdfs-parquet-table-writer.cc:    const string& col_name = 
table_desc_->col_descs()[i + num_clustering_cols].name();
  parquet/parquet-metadata-utils.cc:      const string& name = 
tbl_desc_.col_descs()[table_idx].name();
  scan-node.cc:  VLOG_RPC << "Thread started: " << thread->name();
  scanner-context.cc:              << parent_->scan_node_->hdfs_table()->name()


http://gerrit.cloudera.org:8080/#/c/14660/2//COMMIT_MSG@25
PS2, Line 25: 3. kudu table "test.kudu_table", it would show as
Same here - it looks like there's a pre-existing bug.



--
To view, visit http://gerrit.cloudera.org:8080/14660
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: If5da1112bcf38ae55b89eccfd7c7fad860819a99
Gerrit-Change-Number: 14660
Gerrit-PatchSet: 2
Gerrit-Owner: Xiaomeng Zhang <[email protected]>
Gerrit-Reviewer: Andrew Sherman <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Tim Armstrong <[email protected]>
Gerrit-Reviewer: Xiaomeng Zhang <[email protected]>
Gerrit-Comment-Date: Tue, 19 Nov 2019 01:40:48 +0000
Gerrit-HasComments: Yes

Reply via email to