Search in sources :

Example 1 with PrimaryKeyInfo

use of org.apache.hadoop.hive.ql.metadata.PrimaryKeyInfo in project hive by apache.

the class DDLTask method describeTable.

/**
 * Write the description of a table to a file.
 *
 * @param db
 *          The database in question.
 * @param descTbl
 *          This is the table we're interested in.
 * @return Returns 0 when execution succeeds and above 0 if it fails.
 * @throws HiveException
 *           Throws this exception if an unexpected error occurs.
 * @throws MetaException
 */
private int describeTable(Hive db, DescTableDesc descTbl) throws HiveException, MetaException {
    String colPath = descTbl.getColumnPath();
    String tableName = descTbl.getTableName();
    // describe the table - populate the output stream
    Table tbl = db.getTable(tableName, false);
    if (tbl == null) {
        throw new HiveException(ErrorMsg.INVALID_TABLE, tableName);
    }
    Partition part = null;
    if (descTbl.getPartSpec() != null) {
        part = db.getPartition(tbl, descTbl.getPartSpec(), false);
        if (part == null) {
            throw new HiveException(ErrorMsg.INVALID_PARTITION, StringUtils.join(descTbl.getPartSpec().keySet(), ','), tableName);
        }
        tbl = part.getTable();
    }
    DataOutputStream outStream = getOutputStream(descTbl.getResFile());
    try {
        LOG.debug("DDLTask: got data for {}", tableName);
        List<FieldSchema> cols = null;
        List<ColumnStatisticsObj> colStats = null;
        Deserializer deserializer = tbl.getDeserializer(true);
        if (deserializer instanceof AbstractSerDe) {
            String errorMsgs = ((AbstractSerDe) deserializer).getConfigurationErrors();
            if (errorMsgs != null && !errorMsgs.isEmpty()) {
                throw new SQLException(errorMsgs);
            }
        }
        if (colPath.equals(tableName)) {
            cols = (part == null || tbl.getTableType() == TableType.VIRTUAL_VIEW) ? tbl.getCols() : part.getCols();
            if (!descTbl.isFormatted()) {
                cols.addAll(tbl.getPartCols());
            }
            if (tbl.isPartitioned() && part == null) {
                // No partitioned specified for partitioned table, lets fetch all.
                Map<String, String> tblProps = tbl.getParameters() == null ? new HashMap<String, String>() : tbl.getParameters();
                Map<String, Long> valueMap = new HashMap<>();
                Map<String, Boolean> stateMap = new HashMap<>();
                for (String stat : StatsSetupConst.supportedStats) {
                    valueMap.put(stat, 0L);
                    stateMap.put(stat, true);
                }
                PartitionIterable parts = new PartitionIterable(db, tbl, null, conf.getIntVar(HiveConf.ConfVars.METASTORE_BATCH_RETRIEVE_MAX));
                int numParts = 0;
                for (Partition partition : parts) {
                    Map<String, String> props = partition.getParameters();
                    Boolean state = StatsSetupConst.areBasicStatsUptoDate(props);
                    for (String stat : StatsSetupConst.supportedStats) {
                        stateMap.put(stat, stateMap.get(stat) && state);
                        if (props != null && props.get(stat) != null) {
                            valueMap.put(stat, valueMap.get(stat) + Long.parseLong(props.get(stat)));
                        }
                    }
                    numParts++;
                }
                for (String stat : StatsSetupConst.supportedStats) {
                    StatsSetupConst.setBasicStatsState(tblProps, Boolean.toString(stateMap.get(stat)));
                    tblProps.put(stat, valueMap.get(stat).toString());
                }
                tblProps.put(StatsSetupConst.NUM_PARTITIONS, Integer.toString(numParts));
                tbl.setParameters(tblProps);
            }
        } else {
            if (descTbl.isFormatted()) {
                // when column name is specified in describe table DDL, colPath will
                // will be table_name.column_name
                String colName = colPath.split("\\.")[1];
                String[] dbTab = Utilities.getDbTableName(tableName);
                List<String> colNames = new ArrayList<String>();
                colNames.add(colName.toLowerCase());
                if (null == part) {
                    if (tbl.isPartitioned()) {
                        Map<String, String> tblProps = tbl.getParameters() == null ? new HashMap<String, String>() : tbl.getParameters();
                        if (tbl.isPartitionKey(colNames.get(0))) {
                            FieldSchema partCol = tbl.getPartColByName(colNames.get(0));
                            cols = Collections.singletonList(partCol);
                            PartitionIterable parts = new PartitionIterable(db, tbl, null, conf.getIntVar(HiveConf.ConfVars.METASTORE_BATCH_RETRIEVE_MAX));
                            ColumnInfo ci = new ColumnInfo(partCol.getName(), TypeInfoUtils.getTypeInfoFromTypeString(partCol.getType()), null, false);
                            ColStatistics cs = StatsUtils.getColStatsForPartCol(ci, parts, conf);
                            ColumnStatisticsData data = new ColumnStatisticsData();
                            ColStatistics.Range r = cs.getRange();
                            StatObjectConverter.fillColumnStatisticsData(partCol.getType(), data, r == null ? null : r.minValue, r == null ? null : r.maxValue, r == null ? null : r.minValue, r == null ? null : r.maxValue, r == null ? null : r.minValue.toString(), r == null ? null : r.maxValue.toString(), cs.getNumNulls(), cs.getCountDistint(), null, cs.getAvgColLen(), cs.getAvgColLen(), cs.getNumTrues(), cs.getNumFalses());
                            ColumnStatisticsObj cso = new ColumnStatisticsObj(partCol.getName(), partCol.getType(), data);
                            colStats = Collections.singletonList(cso);
                            StatsSetupConst.setColumnStatsState(tblProps, colNames);
                        } else {
                            cols = Hive.getFieldsFromDeserializer(colPath, deserializer);
                            List<String> parts = db.getPartitionNames(dbTab[0].toLowerCase(), dbTab[1].toLowerCase(), (short) -1);
                            AggrStats aggrStats = db.getAggrColStatsFor(dbTab[0].toLowerCase(), dbTab[1].toLowerCase(), colNames, parts);
                            colStats = aggrStats.getColStats();
                            if (parts.size() == aggrStats.getPartsFound()) {
                                StatsSetupConst.setColumnStatsState(tblProps, colNames);
                            } else {
                                StatsSetupConst.removeColumnStatsState(tblProps, colNames);
                            }
                        }
                        tbl.setParameters(tblProps);
                    } else {
                        cols = Hive.getFieldsFromDeserializer(colPath, deserializer);
                        colStats = db.getTableColumnStatistics(dbTab[0].toLowerCase(), dbTab[1].toLowerCase(), colNames);
                    }
                } else {
                    List<String> partitions = new ArrayList<String>();
                    partitions.add(part.getName());
                    cols = Hive.getFieldsFromDeserializer(colPath, deserializer);
                    colStats = db.getPartitionColumnStatistics(dbTab[0].toLowerCase(), dbTab[1].toLowerCase(), partitions, colNames).get(part.getName());
                }
            } else {
                cols = Hive.getFieldsFromDeserializer(colPath, deserializer);
            }
        }
        PrimaryKeyInfo pkInfo = null;
        ForeignKeyInfo fkInfo = null;
        UniqueConstraint ukInfo = null;
        NotNullConstraint nnInfo = null;
        DefaultConstraint dInfo = null;
        CheckConstraint cInfo = null;
        if (descTbl.isExt() || descTbl.isFormatted()) {
            pkInfo = db.getPrimaryKeys(tbl.getDbName(), tbl.getTableName());
            fkInfo = db.getForeignKeys(tbl.getDbName(), tbl.getTableName());
            ukInfo = db.getUniqueConstraints(tbl.getDbName(), tbl.getTableName());
            nnInfo = db.getNotNullConstraints(tbl.getDbName(), tbl.getTableName());
            dInfo = db.getDefaultConstraints(tbl.getDbName(), tbl.getTableName());
            cInfo = db.getCheckConstraints(tbl.getDbName(), tbl.getTableName());
        }
        fixDecimalColumnTypeName(cols);
        // In case the query is served by HiveServer2, don't pad it with spaces,
        // as HiveServer2 output is consumed by JDBC/ODBC clients.
        boolean isOutputPadded = !SessionState.get().isHiveServerQuery();
        formatter.describeTable(outStream, colPath, tableName, tbl, part, cols, descTbl.isFormatted(), descTbl.isExt(), isOutputPadded, colStats, pkInfo, fkInfo, ukInfo, nnInfo, dInfo, cInfo);
        LOG.debug("DDLTask: written data for {}", tableName);
    } catch (SQLException e) {
        throw new HiveException(e, ErrorMsg.GENERIC_ERROR, tableName);
    } finally {
        IOUtils.closeStream(outStream);
    }
    return 0;
}
Also used : HiveException(org.apache.hadoop.hive.ql.metadata.HiveException) SQLException(java.sql.SQLException) LinkedHashMap(java.util.LinkedHashMap) HashMap(java.util.HashMap) AggrStats(org.apache.hadoop.hive.metastore.api.AggrStats) DataOutputStream(java.io.DataOutputStream) FSDataOutputStream(org.apache.hadoop.fs.FSDataOutputStream) FieldSchema(org.apache.hadoop.hive.metastore.api.FieldSchema) ArrayList(java.util.ArrayList) SQLUniqueConstraint(org.apache.hadoop.hive.metastore.api.SQLUniqueConstraint) UniqueConstraint(org.apache.hadoop.hive.ql.metadata.UniqueConstraint) AbstractSerDe(org.apache.hadoop.hive.serde2.AbstractSerDe) SQLDefaultConstraint(org.apache.hadoop.hive.metastore.api.SQLDefaultConstraint) DefaultConstraint(org.apache.hadoop.hive.ql.metadata.DefaultConstraint) ColumnStatisticsObj(org.apache.hadoop.hive.metastore.api.ColumnStatisticsObj) PrimaryKeyInfo(org.apache.hadoop.hive.ql.metadata.PrimaryKeyInfo) ForeignKeyInfo(org.apache.hadoop.hive.ql.metadata.ForeignKeyInfo) ColStatistics(org.apache.hadoop.hive.ql.plan.ColStatistics) NotNullConstraint(org.apache.hadoop.hive.ql.metadata.NotNullConstraint) SQLNotNullConstraint(org.apache.hadoop.hive.metastore.api.SQLNotNullConstraint) CheckConstraint(org.apache.hadoop.hive.ql.metadata.CheckConstraint) SQLCheckConstraint(org.apache.hadoop.hive.metastore.api.SQLCheckConstraint) Partition(org.apache.hadoop.hive.ql.metadata.Partition) AlterTableExchangePartition(org.apache.hadoop.hive.ql.plan.AlterTableExchangePartition) TextMetaDataTable(org.apache.hadoop.hive.ql.metadata.formatting.TextMetaDataTable) Table(org.apache.hadoop.hive.ql.metadata.Table) SQLUniqueConstraint(org.apache.hadoop.hive.metastore.api.SQLUniqueConstraint) CheckConstraint(org.apache.hadoop.hive.ql.metadata.CheckConstraint) NotNullConstraint(org.apache.hadoop.hive.ql.metadata.NotNullConstraint) SQLCheckConstraint(org.apache.hadoop.hive.metastore.api.SQLCheckConstraint) SQLDefaultConstraint(org.apache.hadoop.hive.metastore.api.SQLDefaultConstraint) DefaultConstraint(org.apache.hadoop.hive.ql.metadata.DefaultConstraint) UniqueConstraint(org.apache.hadoop.hive.ql.metadata.UniqueConstraint) SQLNotNullConstraint(org.apache.hadoop.hive.metastore.api.SQLNotNullConstraint) PartitionIterable(org.apache.hadoop.hive.ql.metadata.PartitionIterable) Deserializer(org.apache.hadoop.hive.serde2.Deserializer) ColumnStatisticsData(org.apache.hadoop.hive.metastore.api.ColumnStatisticsData)

Example 2 with PrimaryKeyInfo

use of org.apache.hadoop.hive.ql.metadata.PrimaryKeyInfo in project hive by apache.

the class RelOptHiveTable method generateKeys.

private Pair<List<ImmutableBitSet>, List<ImmutableBitSet>> generateKeys() {
    final PrimaryKeyInfo primaryKeyInfo = hiveTblMetadata.getPrimaryKeyInfo();
    final UniqueConstraint uniqueKeyInfo = hiveTblMetadata.getUniqueKeyInfo();
    ImmutableList.Builder<ImmutableBitSet> builder = ImmutableList.builder();
    ImmutableList.Builder<ImmutableBitSet> nonNullbuilder = ImmutableList.builder();
    // First PK
    if (primaryKeyInfo != null && !primaryKeyInfo.getColNames().isEmpty()) {
        ImmutableBitSet.Builder keys = ImmutableBitSet.builder();
        for (String pkColName : primaryKeyInfo.getColNames().values()) {
            int pkPos;
            for (pkPos = 0; pkPos < rowType.getFieldNames().size(); pkPos++) {
                String colName = rowType.getFieldNames().get(pkPos);
                if (pkColName.equals(colName)) {
                    break;
                }
            }
            if (pkPos == rowType.getFieldNames().size()) {
                LOG.error("Column for primary key definition " + pkColName + " not found");
            }
            keys.set(pkPos);
        }
        ImmutableBitSet key = keys.build();
        builder.add(key);
        nonNullbuilder.add(key);
    }
    // Then UKs
    if (uniqueKeyInfo != null && !uniqueKeyInfo.getUniqueConstraints().isEmpty()) {
        for (List<UniqueConstraintCol> ukCols : uniqueKeyInfo.getUniqueConstraints().values()) {
            ImmutableBitSet.Builder keys = ImmutableBitSet.builder();
            boolean isNonNullable = true;
            for (UniqueConstraintCol ukCol : ukCols) {
                int ukPos;
                for (ukPos = 0; ukPos < rowType.getFieldNames().size(); ukPos++) {
                    String colName = rowType.getFieldNames().get(ukPos);
                    if (ukCol.colName.equals(colName)) {
                        if (rowType.getFieldList().get(ukPos).getType().isNullable()) {
                            // they should all be nullable
                            isNonNullable = false;
                        }
                        break;
                    }
                }
                if (ukPos == rowType.getFieldNames().size()) {
                    LOG.error("Column for unique constraint definition " + ukCol.colName + " not found");
                }
                keys.set(ukPos);
            }
            ImmutableBitSet key = keys.build();
            builder.add(key);
            if (isNonNullable) {
                nonNullbuilder.add(key);
            }
        }
    }
    return new Pair<>(builder.build(), nonNullbuilder.build());
}
Also used : UniqueConstraintCol(org.apache.hadoop.hive.ql.metadata.UniqueConstraint.UniqueConstraintCol) PrimaryKeyInfo(org.apache.hadoop.hive.ql.metadata.PrimaryKeyInfo) ImmutableBitSet(org.apache.calcite.util.ImmutableBitSet) ImmutableList(com.google.common.collect.ImmutableList) UniqueConstraint(org.apache.hadoop.hive.ql.metadata.UniqueConstraint) UniqueConstraint(org.apache.hadoop.hive.ql.metadata.UniqueConstraint) RelReferentialConstraint(org.apache.calcite.rel.RelReferentialConstraint) IntPair(org.apache.calcite.util.mapping.IntPair) Pair(org.apache.calcite.util.Pair)

Aggregations

PrimaryKeyInfo (org.apache.hadoop.hive.ql.metadata.PrimaryKeyInfo)2 UniqueConstraint (org.apache.hadoop.hive.ql.metadata.UniqueConstraint)2 ImmutableList (com.google.common.collect.ImmutableList)1 DataOutputStream (java.io.DataOutputStream)1 SQLException (java.sql.SQLException)1 ArrayList (java.util.ArrayList)1 HashMap (java.util.HashMap)1 LinkedHashMap (java.util.LinkedHashMap)1 RelReferentialConstraint (org.apache.calcite.rel.RelReferentialConstraint)1 ImmutableBitSet (org.apache.calcite.util.ImmutableBitSet)1 Pair (org.apache.calcite.util.Pair)1 IntPair (org.apache.calcite.util.mapping.IntPair)1 FSDataOutputStream (org.apache.hadoop.fs.FSDataOutputStream)1 AggrStats (org.apache.hadoop.hive.metastore.api.AggrStats)1 ColumnStatisticsData (org.apache.hadoop.hive.metastore.api.ColumnStatisticsData)1 ColumnStatisticsObj (org.apache.hadoop.hive.metastore.api.ColumnStatisticsObj)1 FieldSchema (org.apache.hadoop.hive.metastore.api.FieldSchema)1 SQLCheckConstraint (org.apache.hadoop.hive.metastore.api.SQLCheckConstraint)1 SQLDefaultConstraint (org.apache.hadoop.hive.metastore.api.SQLDefaultConstraint)1 SQLNotNullConstraint (org.apache.hadoop.hive.metastore.api.SQLNotNullConstraint)1