use of org.apache.hadoop.hive.metastore.api.ColumnStatistics in project hive by apache.
the class SessionHiveMetaStoreClient method setPartitionColumnStatistics.
/** {@inheritDoc} */
@Override
public boolean setPartitionColumnStatistics(SetPartitionsStatsRequest request) throws NoSuchObjectException, InvalidObjectException, MetaException, TException, InvalidInputException {
if (request.getColStatsSize() == 1) {
ColumnStatistics colStats = request.getColStatsIterator().next();
ColumnStatisticsDesc desc = colStats.getStatsDesc();
String dbName = desc.getDbName().toLowerCase();
String tableName = desc.getTableName().toLowerCase();
if (getTempTable(dbName, tableName) != null) {
return updateTempTableColumnStats(dbName, tableName, colStats);
}
}
return super.setPartitionColumnStatistics(request);
}
use of org.apache.hadoop.hive.metastore.api.ColumnStatistics in project hive by apache.
the class HBaseReadWrite method getPartitionStatistics.
/**
* Get statistics for a set of partitions
*
* @param dbName name of database table is in
* @param tblName table partitions are in
* @param partNames names of the partitions, used only to set values inside the return stats
* objects
* @param partVals partition values for each partition, needed because this class doesn't know how
* to translate from partName to partVals
* @param colNames column names to fetch stats for. These columns will be fetched for all
* requested partitions
* @return list of ColumnStats, one for each partition for which we found at least one column's
* stats.
* @throws IOException
*/
List<ColumnStatistics> getPartitionStatistics(String dbName, String tblName, List<String> partNames, List<List<String>> partVals, List<String> colNames) throws IOException {
List<ColumnStatistics> statsList = new ArrayList<>(partNames.size());
Map<List<String>, String> valToPartMap = new HashMap<>(partNames.size());
List<Get> gets = new ArrayList<>(partNames.size() * colNames.size());
assert partNames.size() == partVals.size();
byte[][] colNameBytes = new byte[colNames.size()][];
for (int i = 0; i < colNames.size(); i++) {
colNameBytes[i] = HBaseUtils.buildKey(colNames.get(i));
}
for (int i = 0; i < partNames.size(); i++) {
valToPartMap.put(partVals.get(i), partNames.get(i));
byte[] partKey = HBaseUtils.buildPartitionKey(dbName, tblName, HBaseUtils.getPartitionKeyTypes(getTable(dbName, tblName).getPartitionKeys()), partVals.get(i));
Get get = new Get(partKey);
for (byte[] colName : colNameBytes) {
get.addColumn(STATS_CF, colName);
}
gets.add(get);
}
HTableInterface htab = conn.getHBaseTable(PART_TABLE);
Result[] results = htab.get(gets);
for (int i = 0; i < results.length; i++) {
ColumnStatistics colStats = null;
for (int j = 0; j < colNameBytes.length; j++) {
byte[] serializedColStats = results[i].getValue(STATS_CF, colNameBytes[j]);
if (serializedColStats != null) {
if (colStats == null) {
// We initialize this late so that we don't create extras in the case of
// partitions with no stats
colStats = buildColStats(results[i].getRow(), false);
statsList.add(colStats);
}
ColumnStatisticsObj cso = HBaseUtils.deserializeStatsForOneColumn(colStats, serializedColStats);
cso.setColName(colNames.get(j));
colStats.addToStatsObj(cso);
}
}
}
return statsList;
}
use of org.apache.hadoop.hive.metastore.api.ColumnStatistics in project hive by apache.
the class HBaseReadWrite method getTableStatistics.
/**
* Get statistics for a table
*
* @param dbName name of database table is in
* @param tblName name of table
* @param colNames list of column names to get statistics for
* @return column statistics for indicated table
* @throws IOException
*/
ColumnStatistics getTableStatistics(String dbName, String tblName, List<String> colNames) throws IOException {
byte[] tabKey = HBaseUtils.buildKey(dbName, tblName);
ColumnStatistics tableStats = new ColumnStatistics();
ColumnStatisticsDesc statsDesc = new ColumnStatisticsDesc();
statsDesc.setIsTblLevel(true);
statsDesc.setDbName(dbName);
statsDesc.setTableName(tblName);
tableStats.setStatsDesc(statsDesc);
byte[][] colKeys = new byte[colNames.size()][];
for (int i = 0; i < colKeys.length; i++) {
colKeys[i] = HBaseUtils.buildKey(colNames.get(i));
}
Result result = read(TABLE_TABLE, tabKey, STATS_CF, colKeys);
for (int i = 0; i < colKeys.length; i++) {
byte[] serializedColStats = result.getValue(STATS_CF, colKeys[i]);
if (serializedColStats == null) {
// There were no stats for this column, so skip it
continue;
}
ColumnStatisticsObj obj = HBaseUtils.deserializeStatsForOneColumn(tableStats, serializedColStats);
obj.setColName(colNames.get(i));
tableStats.addToStatsObj(obj);
}
return tableStats;
}
use of org.apache.hadoop.hive.metastore.api.ColumnStatistics in project hive by apache.
the class HBaseReadWrite method printOnePartition.
private String printOnePartition(Result result) throws IOException, TException {
byte[] key = result.getRow();
HBaseUtils.StorageDescriptorParts sdParts = HBaseUtils.deserializePartition(key, result.getValue(CATALOG_CF, CATALOG_COL), this);
StringBuilder builder = new StringBuilder();
builder.append(dumpThriftObject(sdParts.containingPartition)).append(" sdHash: ").append(Base64.encodeBase64URLSafeString(sdParts.sdHash)).append(" stats:");
NavigableMap<byte[], byte[]> statsCols = result.getFamilyMap(STATS_CF);
for (Map.Entry<byte[], byte[]> statsCol : statsCols.entrySet()) {
builder.append(" column ").append(new String(statsCol.getKey(), HBaseUtils.ENCODING)).append(": ");
ColumnStatistics pcs = buildColStats(key, false);
ColumnStatisticsObj cso = HBaseUtils.deserializeStatsForOneColumn(pcs, statsCol.getValue());
builder.append(dumpThriftObject(cso));
}
return builder.toString();
}
Aggregations