Examples with StringColumnStatsDataInspector org.apache.hadoop.hive.metastore.columnstats.cache.StringColumnStatsDataInspector used on opensource projects

Example 6 with StringColumnStatsDataInspector

use of org.apache.hadoop.hive.metastore.columnstats.cache.StringColumnStatsDataInspector in project hive by apache.

the class ColumnStatsMergerFactory method newColumnStaticsObj.

public static ColumnStatisticsObj newColumnStaticsObj(final String colName, final String colType, final _Fields type) {
    final ColumnStatisticsObj cso = new ColumnStatisticsObj();
    final ColumnStatisticsData csd = new ColumnStatisticsData();
    Objects.requireNonNull(colName, "Column name cannot be null");
    Objects.requireNonNull(colType, "Column type cannot be null");
    Objects.requireNonNull(type, "Field type cannot be null");
    switch(type) {
        case BOOLEAN_STATS:
            csd.setBooleanStats(new BooleanColumnStatsData());
            break;
        case LONG_STATS:
            csd.setLongStats(new LongColumnStatsDataInspector());
            break;
        case DOUBLE_STATS:
            csd.setDoubleStats(new DoubleColumnStatsDataInspector());
            break;
        case STRING_STATS:
            csd.setStringStats(new StringColumnStatsDataInspector());
            break;
        case BINARY_STATS:
            csd.setBinaryStats(new BinaryColumnStatsData());
            break;
        case DECIMAL_STATS:
            csd.setDecimalStats(new DecimalColumnStatsDataInspector());
            break;
        case DATE_STATS:
            csd.setDateStats(new DateColumnStatsDataInspector());
            break;
        case TIMESTAMP_STATS:
            csd.setTimestampStats(new TimestampColumnStatsDataInspector());
            break;
        default:
            throw new IllegalArgumentException("Unknown stats type: " + type);
    }
    cso.setColName(colName);
    cso.setColType(colType);
    cso.setStatsData(csd);
    return cso;
}

Example 7 with StringColumnStatsDataInspector

use of org.apache.hadoop.hive.metastore.columnstats.cache.StringColumnStatsDataInspector in project hive by apache.

the class StringColumnStatsAggregator method extrapolate.

@Override
public void extrapolate(ColumnStatisticsData extrapolateData, int numParts, int numPartsWithStats, Map<String, Double> adjustedIndexMap, Map<String, ColumnStatisticsData> adjustedStatsMap, double densityAvg) {
    int rightBorderInd = numParts;
    StringColumnStatsDataInspector extrapolateStringData = new StringColumnStatsDataInspector();
    Map<String, StringColumnStatsData> extractedAdjustedStatsMap = new HashMap<>();
    for (Map.Entry<String, ColumnStatisticsData> entry : adjustedStatsMap.entrySet()) {
        extractedAdjustedStatsMap.put(entry.getKey(), entry.getValue().getStringStats());
    }
    List<Map.Entry<String, StringColumnStatsData>> list = new LinkedList<>(extractedAdjustedStatsMap.entrySet());
    // get the avgLen
    Collections.sort(list, new Comparator<Map.Entry<String, StringColumnStatsData>>() {

        @Override
        public int compare(Map.Entry<String, StringColumnStatsData> o1, Map.Entry<String, StringColumnStatsData> o2) {
            return Double.compare(o1.getValue().getAvgColLen(), o2.getValue().getAvgColLen());
        }
    });
    double minInd = adjustedIndexMap.get(list.get(0).getKey());
    double maxInd = adjustedIndexMap.get(list.get(list.size() - 1).getKey());
    double avgColLen = 0;
    double min = list.get(0).getValue().getAvgColLen();
    double max = list.get(list.size() - 1).getValue().getAvgColLen();
    if (minInd == maxInd) {
        avgColLen = min;
    } else if (minInd < maxInd) {
        // right border is the max
        avgColLen = (min + (max - min) * (rightBorderInd - minInd) / (maxInd - minInd));
    } else {
        // left border is the max
        avgColLen = (min + (max - min) * minInd / (minInd - maxInd));
    }
    // get the maxLen
    Collections.sort(list, new Comparator<Map.Entry<String, StringColumnStatsData>>() {

        @Override
        public int compare(Map.Entry<String, StringColumnStatsData> o1, Map.Entry<String, StringColumnStatsData> o2) {
            return Long.compare(o1.getValue().getMaxColLen(), o2.getValue().getMaxColLen());
        }
    });
    minInd = adjustedIndexMap.get(list.get(0).getKey());
    maxInd = adjustedIndexMap.get(list.get(list.size() - 1).getKey());
    double maxColLen = 0;
    min = list.get(0).getValue().getAvgColLen();
    max = list.get(list.size() - 1).getValue().getAvgColLen();
    if (minInd == maxInd) {
        maxColLen = min;
    } else if (minInd < maxInd) {
        // right border is the max
        maxColLen = (min + (max - min) * (rightBorderInd - minInd) / (maxInd - minInd));
    } else {
        // left border is the max
        maxColLen = (min + (max - min) * minInd / (minInd - maxInd));
    }
    // get the #nulls
    long numNulls = 0;
    for (Map.Entry<String, StringColumnStatsData> entry : extractedAdjustedStatsMap.entrySet()) {
        numNulls += entry.getValue().getNumNulls();
    }
    // we scale up sumNulls based on the number of partitions
    numNulls = numNulls * numParts / numPartsWithStats;
    // get the ndv
    long ndv = 0;
    Collections.sort(list, new Comparator<Map.Entry<String, StringColumnStatsData>>() {

        @Override
        public int compare(Map.Entry<String, StringColumnStatsData> o1, Map.Entry<String, StringColumnStatsData> o2) {
            return Long.compare(o1.getValue().getNumDVs(), o2.getValue().getNumDVs());
        }
    });
    minInd = adjustedIndexMap.get(list.get(0).getKey());
    maxInd = adjustedIndexMap.get(list.get(list.size() - 1).getKey());
    min = list.get(0).getValue().getNumDVs();
    max = list.get(list.size() - 1).getValue().getNumDVs();
    if (minInd == maxInd) {
        ndv = (long) min;
    } else if (minInd < maxInd) {
        // right border is the max
        ndv = (long) (min + (max - min) * (rightBorderInd - minInd) / (maxInd - minInd));
    } else {
        // left border is the max
        ndv = (long) (min + (max - min) * minInd / (minInd - maxInd));
    }
    extrapolateStringData.setAvgColLen(avgColLen);
    extrapolateStringData.setMaxColLen((long) maxColLen);
    extrapolateStringData.setNumNulls(numNulls);
    extrapolateStringData.setNumDVs(ndv);
    extrapolateData.setStringStats(extrapolateStringData);
}

Example 8 with StringColumnStatsDataInspector

use of org.apache.hadoop.hive.metastore.columnstats.cache.StringColumnStatsDataInspector in project hive by apache.

the class ColumnStatisticsObjTranslator method unpackPrimitiveObject.

private static void unpackPrimitiveObject(ObjectInspector oi, Object o, String fieldName, ColumnStatisticsObj statsObj) throws UnsupportedDoubleException {
    if (o == null) {
        return;
    }
    // First infer the type of object
    if (fieldName.equals("columntype")) {
        PrimitiveObjectInspector poi = (PrimitiveObjectInspector) oi;
        String s = ((StringObjectInspector) poi).getPrimitiveJavaObject(o);
        ColumnStatisticsData statsData = new ColumnStatisticsData();
        if (s.equalsIgnoreCase("long")) {
            LongColumnStatsDataInspector longStats = new LongColumnStatsDataInspector();
            statsData.setLongStats(longStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase("double")) {
            DoubleColumnStatsDataInspector doubleStats = new DoubleColumnStatsDataInspector();
            statsData.setDoubleStats(doubleStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase("string")) {
            StringColumnStatsDataInspector stringStats = new StringColumnStatsDataInspector();
            statsData.setStringStats(stringStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase("boolean")) {
            BooleanColumnStatsData booleanStats = new BooleanColumnStatsData();
            statsData.setBooleanStats(booleanStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase("binary")) {
            BinaryColumnStatsData binaryStats = new BinaryColumnStatsData();
            statsData.setBinaryStats(binaryStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase("decimal")) {
            DecimalColumnStatsDataInspector decimalStats = new DecimalColumnStatsDataInspector();
            statsData.setDecimalStats(decimalStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase("date")) {
            DateColumnStatsDataInspector dateStats = new DateColumnStatsDataInspector();
            statsData.setDateStats(dateStats);
            statsObj.setStatsData(statsData);
        }
    } else {
        // invoke the right unpack method depending on data type of the column
        if (statsObj.getStatsData().isSetBooleanStats()) {
            unpackBooleanStats(oi, o, fieldName, statsObj);
        } else if (statsObj.getStatsData().isSetLongStats()) {
            unpackLongStats(oi, o, fieldName, statsObj);
        } else if (statsObj.getStatsData().isSetDoubleStats()) {
            unpackDoubleStats(oi, o, fieldName, statsObj);
        } else if (statsObj.getStatsData().isSetStringStats()) {
            unpackStringStats(oi, o, fieldName, statsObj);
        } else if (statsObj.getStatsData().isSetBinaryStats()) {
            unpackBinaryStats(oi, o, fieldName, statsObj);
        } else if (statsObj.getStatsData().isSetDecimalStats()) {
            unpackDecimalStats(oi, o, fieldName, statsObj);
        } else if (statsObj.getStatsData().isSetDateStats()) {
            unpackDateStats(oi, o, fieldName, statsObj);
        }
    }
}

Example 9 with StringColumnStatsDataInspector

use of org.apache.hadoop.hive.metastore.columnstats.cache.StringColumnStatsDataInspector in project hive by apache.

the class TestCachedStore method testTableColStatsOps.

// @Test
public void testTableColStatsOps() throws Exception {
    // Add a db via ObjectStore
    String dbName = "testTableColStatsOps";
    String dbOwner = "user1";
    Database db = createTestDb(dbName, dbOwner);
    objectStore.createDatabase(db);
    db = objectStore.getDatabase(dbName);
    // Add a table via ObjectStore
    final String tblName = "tbl";
    final String tblOwner = "user1";
    final FieldSchema col1 = new FieldSchema("col1", "int", "integer column");
    // Stats values for col1
    long col1LowVal = 5;
    long col1HighVal = 500;
    long col1Nulls = 10;
    long col1DV = 20;
    final FieldSchema col2 = new FieldSchema("col2", "string", "string column");
    // Stats values for col2
    long col2MaxColLen = 100;
    double col2AvgColLen = 45.5;
    long col2Nulls = 5;
    long col2DV = 40;
    final FieldSchema col3 = new FieldSchema("col3", "boolean", "boolean column");
    // Stats values for col3
    long col3NumTrues = 100;
    long col3NumFalses = 30;
    long col3Nulls = 10;
    final List<FieldSchema> cols = new ArrayList<>();
    cols.add(col1);
    cols.add(col2);
    cols.add(col3);
    FieldSchema ptnCol1 = new FieldSchema("part1", "string", "string partition column");
    List<FieldSchema> ptnCols = new ArrayList<FieldSchema>();
    ptnCols.add(ptnCol1);
    Table tbl = createTestTbl(dbName, tblName, tblOwner, cols, ptnCols);
    objectStore.createTable(tbl);
    tbl = objectStore.getTable(dbName, tblName);
    // Add ColumnStatistics for tbl to metastore DB via ObjectStore
    ColumnStatistics stats = new ColumnStatistics();
    ColumnStatisticsDesc statsDesc = new ColumnStatisticsDesc(true, dbName, tblName);
    List<ColumnStatisticsObj> colStatObjs = new ArrayList<>();
    // Col1
    ColumnStatisticsData data1 = new ColumnStatisticsData();
    ColumnStatisticsObj col1Stats = new ColumnStatisticsObj(col1.getName(), col1.getType(), data1);
    LongColumnStatsDataInspector longStats = new LongColumnStatsDataInspector();
    longStats.setLowValue(col1LowVal);
    longStats.setHighValue(col1HighVal);
    longStats.setNumNulls(col1Nulls);
    longStats.setNumDVs(col1DV);
    data1.setLongStats(longStats);
    colStatObjs.add(col1Stats);
    // Col2
    ColumnStatisticsData data2 = new ColumnStatisticsData();
    ColumnStatisticsObj col2Stats = new ColumnStatisticsObj(col2.getName(), col2.getType(), data2);
    StringColumnStatsDataInspector stringStats = new StringColumnStatsDataInspector();
    stringStats.setMaxColLen(col2MaxColLen);
    stringStats.setAvgColLen(col2AvgColLen);
    stringStats.setNumNulls(col2Nulls);
    stringStats.setNumDVs(col2DV);
    data2.setStringStats(stringStats);
    colStatObjs.add(col2Stats);
    // Col3
    ColumnStatisticsData data3 = new ColumnStatisticsData();
    ColumnStatisticsObj col3Stats = new ColumnStatisticsObj(col3.getName(), col3.getType(), data3);
    BooleanColumnStatsData boolStats = new BooleanColumnStatsData();
    boolStats.setNumTrues(col3NumTrues);
    boolStats.setNumFalses(col3NumFalses);
    boolStats.setNumNulls(col3Nulls);
    data3.setBooleanStats(boolStats);
    colStatObjs.add(col3Stats);
    stats.setStatsDesc(statsDesc);
    stats.setStatsObj(colStatObjs);
    // Save to DB
    objectStore.updateTableColumnStatistics(stats);
    // Prewarm CachedStore
    CachedStore.setCachePrewarmedState(false);
    CachedStore.prewarm(objectStore);
    // Read table stats via CachedStore
    ColumnStatistics newStats = cachedStore.getTableColumnStatistics(dbName, tblName, Arrays.asList(col1.getName(), col2.getName(), col3.getName()));
    Assert.assertEquals(stats, newStats);
    // Clean up
    objectStore.dropTable(dbName, tblName);
    objectStore.dropDatabase(dbName);
    sharedCache.getDatabaseCache().clear();
    sharedCache.getTableCache().clear();
    sharedCache.getSdCache().clear();
}

Example 10 with StringColumnStatsDataInspector

use of org.apache.hadoop.hive.metastore.columnstats.cache.StringColumnStatsDataInspector in project hive by apache.

the class ColumnStatisticsObjTranslator method unpackPrimitiveObject.

private static void unpackPrimitiveObject(ObjectInspector oi, Object o, ColumnStatsField csf, ColumnStatisticsObj statsObj) throws UnsupportedDoubleException {
    if (o == null) {
        return;
    }
    // First infer the type of object
    if (csf == ColumnStatsField.COLUMN_STATS_TYPE) {
        PrimitiveObjectInspector poi = (PrimitiveObjectInspector) oi;
        String s = ((StringObjectInspector) poi).getPrimitiveJavaObject(o);
        ColumnStatisticsData statsData = new ColumnStatisticsData();
        if (s.equalsIgnoreCase(ColumnStatsType.LONG.toString())) {
            LongColumnStatsDataInspector longStats = new LongColumnStatsDataInspector();
            statsData.setLongStats(longStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase(ColumnStatsType.DOUBLE.toString())) {
            DoubleColumnStatsDataInspector doubleStats = new DoubleColumnStatsDataInspector();
            statsData.setDoubleStats(doubleStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase(ColumnStatsType.STRING.toString())) {
            StringColumnStatsDataInspector stringStats = new StringColumnStatsDataInspector();
            statsData.setStringStats(stringStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase(ColumnStatsType.BOOLEAN.toString())) {
            BooleanColumnStatsData booleanStats = new BooleanColumnStatsData();
            statsData.setBooleanStats(booleanStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase(ColumnStatsType.BINARY.toString())) {
            BinaryColumnStatsData binaryStats = new BinaryColumnStatsData();
            statsData.setBinaryStats(binaryStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase(ColumnStatsType.DECIMAL.toString())) {
            DecimalColumnStatsDataInspector decimalStats = new DecimalColumnStatsDataInspector();
            statsData.setDecimalStats(decimalStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase(ColumnStatsType.DATE.toString())) {
            DateColumnStatsDataInspector dateStats = new DateColumnStatsDataInspector();
            statsData.setDateStats(dateStats);
            statsObj.setStatsData(statsData);
        } else if (s.equalsIgnoreCase(ColumnStatsType.TIMESTAMP.toString())) {
            TimestampColumnStatsDataInspector timestampStats = new TimestampColumnStatsDataInspector();
            statsData.setTimestampStats(timestampStats);
            statsObj.setStatsData(statsData);
        }
    } else {
        // invoke the right unpack method depending on data type of the column
        if (statsObj.getStatsData().isSetBooleanStats()) {
            unpackBooleanStats(oi, o, csf, statsObj);
        } else if (statsObj.getStatsData().isSetLongStats()) {
            unpackLongStats(oi, o, csf, statsObj);
        } else if (statsObj.getStatsData().isSetDoubleStats()) {
            unpackDoubleStats(oi, o, csf, statsObj);
        } else if (statsObj.getStatsData().isSetStringStats()) {
            unpackStringStats(oi, o, csf, statsObj);
        } else if (statsObj.getStatsData().isSetBinaryStats()) {
            unpackBinaryStats(oi, o, csf, statsObj);
        } else if (statsObj.getStatsData().isSetDecimalStats()) {
            unpackDecimalStats(oi, o, csf, statsObj);
        } else if (statsObj.getStatsData().isSetDateStats()) {
            unpackDateStats(oi, o, csf, statsObj);
        } else if (statsObj.getStatsData().isSetTimestampStats()) {
            unpackTimestampStats(oi, o, csf, statsObj);
        }
    }
}