Search in sources :

Example 61 with ColumnStatistics

use of org.apache.hadoop.hive.metastore.api.ColumnStatistics in project hive by apache.

the class TestHBaseAggrStatsCacheIntegration method invalidation.

@Test
public void invalidation() throws Exception {
    try {
        String dbName = "default";
        String tableName = "invalidation";
        List<String> partVals1 = Arrays.asList("today");
        List<String> partVals2 = Arrays.asList("yesterday");
        List<String> partVals3 = Arrays.asList("tomorrow");
        long now = System.currentTimeMillis();
        List<FieldSchema> cols = new ArrayList<>();
        cols.add(new FieldSchema("col1", "boolean", "nocomment"));
        SerDeInfo serde = new SerDeInfo("serde", "seriallib", null);
        StorageDescriptor sd = new StorageDescriptor(cols, "file:/tmp", "input", "output", false, 0, serde, null, null, Collections.<String, String>emptyMap());
        List<FieldSchema> partCols = new ArrayList<>();
        partCols.add(new FieldSchema("ds", "string", ""));
        Table table = new Table(tableName, dbName, "me", (int) now, (int) now, 0, sd, partCols, Collections.<String, String>emptyMap(), null, null, null);
        store.createTable(table);
        for (List<String> partVals : Arrays.asList(partVals1, partVals2, partVals3)) {
            StorageDescriptor psd = new StorageDescriptor(sd);
            psd.setLocation("file:/tmp/default/invalidation/ds=" + partVals.get(0));
            Partition part = new Partition(partVals, dbName, tableName, (int) now, (int) now, psd, Collections.<String, String>emptyMap());
            store.addPartition(part);
            ColumnStatistics cs = new ColumnStatistics();
            ColumnStatisticsDesc desc = new ColumnStatisticsDesc(false, dbName, tableName);
            desc.setLastAnalyzed(now);
            desc.setPartName("ds=" + partVals.get(0));
            cs.setStatsDesc(desc);
            ColumnStatisticsObj obj = new ColumnStatisticsObj();
            obj.setColName("col1");
            obj.setColType("boolean");
            ColumnStatisticsData data = new ColumnStatisticsData();
            BooleanColumnStatsData bcsd = new BooleanColumnStatsData();
            bcsd.setNumFalses(10);
            bcsd.setNumTrues(20);
            bcsd.setNumNulls(30);
            data.setBooleanStats(bcsd);
            obj.setStatsData(data);
            cs.addToStatsObj(obj);
            store.updatePartitionColumnStatistics(cs, partVals);
        }
        Checker statChecker = new Checker() {

            @Override
            public void checkStats(AggrStats aggrStats) throws Exception {
                Assert.assertEquals(2, aggrStats.getPartsFound());
                Assert.assertEquals(1, aggrStats.getColStatsSize());
                ColumnStatisticsObj cso = aggrStats.getColStats().get(0);
                Assert.assertEquals("col1", cso.getColName());
                Assert.assertEquals("boolean", cso.getColType());
                BooleanColumnStatsData bcsd = cso.getStatsData().getBooleanStats();
                Assert.assertEquals(20, bcsd.getNumFalses());
                Assert.assertEquals(40, bcsd.getNumTrues());
                Assert.assertEquals(60, bcsd.getNumNulls());
            }
        };
        AggrStats aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=today", "ds=yesterday"), Arrays.asList("col1"));
        statChecker.checkStats(aggrStats);
        // Check that we had to build it from the stats
        Assert.assertEquals(0, store.backdoor().getStatsCache().hbaseHits.getCnt());
        Assert.assertEquals(1, store.backdoor().getStatsCache().totalGets.getCnt());
        Assert.assertEquals(1, store.backdoor().getStatsCache().misses.getCnt());
        // Call again, this time it should come from memory.  Also, reverse the name order this time
        // to assure that we still hit.
        aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=yesterday", "ds=today"), Arrays.asList("col1"));
        statChecker.checkStats(aggrStats);
        Assert.assertEquals(0, store.backdoor().getStatsCache().hbaseHits.getCnt());
        Assert.assertEquals(2, store.backdoor().getStatsCache().totalGets.getCnt());
        Assert.assertEquals(1, store.backdoor().getStatsCache().misses.getCnt());
        // Now call a different combination to get it in memory too
        aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=tomorrow", "ds=today"), Arrays.asList("col1"));
        statChecker.checkStats(aggrStats);
        Assert.assertEquals(0, store.backdoor().getStatsCache().hbaseHits.getCnt());
        Assert.assertEquals(3, store.backdoor().getStatsCache().totalGets.getCnt());
        Assert.assertEquals(2, store.backdoor().getStatsCache().misses.getCnt());
        aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=tomorrow", "ds=today"), Arrays.asList("col1"));
        statChecker.checkStats(aggrStats);
        Assert.assertEquals(0, store.backdoor().getStatsCache().hbaseHits.getCnt());
        Assert.assertEquals(4, store.backdoor().getStatsCache().totalGets.getCnt());
        Assert.assertEquals(2, store.backdoor().getStatsCache().misses.getCnt());
        // wake the invalidator and check again to make sure it isn't too aggressive about
        // removing our stuff.
        store.backdoor().getStatsCache().wakeInvalidator();
        aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=tomorrow", "ds=today"), Arrays.asList("col1"));
        statChecker.checkStats(aggrStats);
        Assert.assertEquals(0, store.backdoor().getStatsCache().hbaseHits.getCnt());
        Assert.assertEquals(5, store.backdoor().getStatsCache().totalGets.getCnt());
        Assert.assertEquals(2, store.backdoor().getStatsCache().misses.getCnt());
        // Update statistics for 'tomorrow'
        ColumnStatistics cs = new ColumnStatistics();
        ColumnStatisticsDesc desc = new ColumnStatisticsDesc(false, dbName, tableName);
        desc.setLastAnalyzed(now);
        desc.setPartName("ds=" + partVals3.get(0));
        cs.setStatsDesc(desc);
        ColumnStatisticsObj obj = new ColumnStatisticsObj();
        obj.setColName("col1");
        obj.setColType("boolean");
        ColumnStatisticsData data = new ColumnStatisticsData();
        BooleanColumnStatsData bcsd = new BooleanColumnStatsData();
        bcsd.setNumFalses(100);
        bcsd.setNumTrues(200);
        bcsd.setNumNulls(300);
        data.setBooleanStats(bcsd);
        obj.setStatsData(data);
        cs.addToStatsObj(obj);
        Checker afterUpdate = new Checker() {

            @Override
            public void checkStats(AggrStats aggrStats) throws Exception {
                Assert.assertEquals(2, aggrStats.getPartsFound());
                Assert.assertEquals(1, aggrStats.getColStatsSize());
                ColumnStatisticsObj cso = aggrStats.getColStats().get(0);
                Assert.assertEquals("col1", cso.getColName());
                Assert.assertEquals("boolean", cso.getColType());
                BooleanColumnStatsData bcsd = cso.getStatsData().getBooleanStats();
                Assert.assertEquals(110, bcsd.getNumFalses());
                Assert.assertEquals(220, bcsd.getNumTrues());
                Assert.assertEquals(330, bcsd.getNumNulls());
            }
        };
        store.updatePartitionColumnStatistics(cs, partVals3);
        store.backdoor().getStatsCache().setRunInvalidatorEvery(100);
        store.backdoor().getStatsCache().wakeInvalidator();
        aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=tomorrow", "ds=today"), Arrays.asList("col1"));
        afterUpdate.checkStats(aggrStats);
        // Check that we missed, which means this aggregate was dropped from the cache.
        Assert.assertEquals(0, store.backdoor().getStatsCache().hbaseHits.getCnt());
        Assert.assertEquals(6, store.backdoor().getStatsCache().totalGets.getCnt());
        Assert.assertEquals(3, store.backdoor().getStatsCache().misses.getCnt());
        // Check that our other aggregate is still in the cache.
        aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=yesterday", "ds=today"), Arrays.asList("col1"));
        statChecker.checkStats(aggrStats);
        Assert.assertEquals(0, store.backdoor().getStatsCache().hbaseHits.getCnt());
        Assert.assertEquals(7, store.backdoor().getStatsCache().totalGets.getCnt());
        Assert.assertEquals(3, store.backdoor().getStatsCache().misses.getCnt());
        // Drop 'yesterday', so our first aggregate should be dumped from memory and hbase
        store.dropPartition(dbName, tableName, partVals2);
        store.backdoor().getStatsCache().wakeInvalidator();
        aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=yesterday", "ds=today"), Arrays.asList("col1"));
        new Checker() {

            @Override
            public void checkStats(AggrStats aggrStats) throws Exception {
                Assert.assertEquals(1, aggrStats.getPartsFound());
                Assert.assertEquals(1, aggrStats.getColStatsSize());
                ColumnStatisticsObj cso = aggrStats.getColStats().get(0);
                Assert.assertEquals("col1", cso.getColName());
                Assert.assertEquals("boolean", cso.getColType());
                BooleanColumnStatsData bcsd = cso.getStatsData().getBooleanStats();
                Assert.assertEquals(10, bcsd.getNumFalses());
                Assert.assertEquals(20, bcsd.getNumTrues());
                Assert.assertEquals(30, bcsd.getNumNulls());
            }
        }.checkStats(aggrStats);
        // Check that we missed, which means this aggregate was dropped from the cache.
        Assert.assertEquals(0, store.backdoor().getStatsCache().hbaseHits.getCnt());
        Assert.assertEquals(8, store.backdoor().getStatsCache().totalGets.getCnt());
        Assert.assertEquals(4, store.backdoor().getStatsCache().misses.getCnt());
        // Check that our other aggregate is still in the cache.
        aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=tomorrow", "ds=today"), Arrays.asList("col1"));
        afterUpdate.checkStats(aggrStats);
        Assert.assertEquals(0, store.backdoor().getStatsCache().hbaseHits.getCnt());
        Assert.assertEquals(9, store.backdoor().getStatsCache().totalGets.getCnt());
        Assert.assertEquals(4, store.backdoor().getStatsCache().misses.getCnt());
    } finally {
        store.backdoor().getStatsCache().setRunInvalidatorEvery(5000);
        store.backdoor().getStatsCache().setMaxTimeInCache(500000);
        store.backdoor().getStatsCache().wakeInvalidator();
    }
}
Also used : ColumnStatistics(org.apache.hadoop.hive.metastore.api.ColumnStatistics) BooleanColumnStatsData(org.apache.hadoop.hive.metastore.api.BooleanColumnStatsData) Partition(org.apache.hadoop.hive.metastore.api.Partition) Table(org.apache.hadoop.hive.metastore.api.Table) AggrStats(org.apache.hadoop.hive.metastore.api.AggrStats) FieldSchema(org.apache.hadoop.hive.metastore.api.FieldSchema) SerDeInfo(org.apache.hadoop.hive.metastore.api.SerDeInfo) ArrayList(java.util.ArrayList) StorageDescriptor(org.apache.hadoop.hive.metastore.api.StorageDescriptor) ExpectedException(org.junit.rules.ExpectedException) IOException(java.io.IOException) ColumnStatisticsObj(org.apache.hadoop.hive.metastore.api.ColumnStatisticsObj) ColumnStatisticsDesc(org.apache.hadoop.hive.metastore.api.ColumnStatisticsDesc) ColumnStatisticsData(org.apache.hadoop.hive.metastore.api.ColumnStatisticsData) Test(org.junit.Test)

Example 62 with ColumnStatistics

use of org.apache.hadoop.hive.metastore.api.ColumnStatistics in project hive by apache.

the class TestHBaseAggrStatsCacheIntegration method altersInvalidation.

@Test
public void altersInvalidation() throws Exception {
    try {
        String dbName = "default";
        String tableName = "asi";
        List<String> partVals1 = Arrays.asList("today");
        List<String> partVals2 = Arrays.asList("yesterday");
        List<String> partVals3 = Arrays.asList("tomorrow");
        long now = System.currentTimeMillis();
        List<FieldSchema> cols = new ArrayList<>();
        cols.add(new FieldSchema("col1", "boolean", "nocomment"));
        SerDeInfo serde = new SerDeInfo("serde", "seriallib", null);
        StorageDescriptor sd = new StorageDescriptor(cols, "file:/tmp", "input", "output", false, 0, serde, null, null, Collections.<String, String>emptyMap());
        List<FieldSchema> partCols = new ArrayList<>();
        partCols.add(new FieldSchema("ds", "string", ""));
        Table table = new Table(tableName, dbName, "me", (int) now, (int) now, 0, sd, partCols, Collections.<String, String>emptyMap(), null, null, null);
        store.createTable(table);
        Partition[] partitions = new Partition[3];
        int partnum = 0;
        for (List<String> partVals : Arrays.asList(partVals1, partVals2, partVals3)) {
            StorageDescriptor psd = new StorageDescriptor(sd);
            psd.setLocation("file:/tmp/default/invalidation/ds=" + partVals.get(0));
            Partition part = new Partition(partVals, dbName, tableName, (int) now, (int) now, psd, Collections.<String, String>emptyMap());
            partitions[partnum++] = part;
            store.addPartition(part);
            ColumnStatistics cs = new ColumnStatistics();
            ColumnStatisticsDesc desc = new ColumnStatisticsDesc(false, dbName, tableName);
            desc.setLastAnalyzed(now);
            desc.setPartName("ds=" + partVals.get(0));
            cs.setStatsDesc(desc);
            ColumnStatisticsObj obj = new ColumnStatisticsObj();
            obj.setColName("col1");
            obj.setColType("boolean");
            ColumnStatisticsData data = new ColumnStatisticsData();
            BooleanColumnStatsData bcsd = new BooleanColumnStatsData();
            bcsd.setNumFalses(10);
            bcsd.setNumTrues(20);
            bcsd.setNumNulls(30);
            data.setBooleanStats(bcsd);
            obj.setStatsData(data);
            cs.addToStatsObj(obj);
            store.updatePartitionColumnStatistics(cs, partVals);
        }
        AggrStats aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=today", "ds=tomorrow"), Arrays.asList("col1"));
        aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=today", "ds=yesterday"), Arrays.asList("col1"));
        // Check that we had to build it from the stats
        Assert.assertEquals(0, store.backdoor().getStatsCache().hbaseHits.getCnt());
        Assert.assertEquals(2, store.backdoor().getStatsCache().totalGets.getCnt());
        Assert.assertEquals(2, store.backdoor().getStatsCache().misses.getCnt());
        // wake the invalidator and check again to make sure it isn't too aggressive about
        // removing our stuff.
        store.backdoor().getStatsCache().wakeInvalidator();
        Partition[] newParts = new Partition[2];
        newParts[0] = new Partition(partitions[0]);
        newParts[0].setLastAccessTime((int) System.currentTimeMillis());
        newParts[1] = new Partition(partitions[2]);
        newParts[1].setLastAccessTime((int) System.currentTimeMillis());
        store.alterPartitions(dbName, tableName, Arrays.asList(partVals1, partVals3), Arrays.asList(newParts));
        store.backdoor().getStatsCache().setRunInvalidatorEvery(100);
        store.backdoor().getStatsCache().wakeInvalidator();
        aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=tomorrow", "ds=today"), Arrays.asList("col1"));
        // Check that we missed, which means this aggregate was dropped from the cache.
        Assert.assertEquals(0, store.backdoor().getStatsCache().hbaseHits.getCnt());
        Assert.assertEquals(3, store.backdoor().getStatsCache().totalGets.getCnt());
        Assert.assertEquals(3, store.backdoor().getStatsCache().misses.getCnt());
        // Check that our other aggregate got dropped too
        aggrStats = store.get_aggr_stats_for(dbName, tableName, Arrays.asList("ds=yesterday", "ds=today"), Arrays.asList("col1"));
        Assert.assertEquals(0, store.backdoor().getStatsCache().hbaseHits.getCnt());
        Assert.assertEquals(4, store.backdoor().getStatsCache().totalGets.getCnt());
        Assert.assertEquals(4, store.backdoor().getStatsCache().misses.getCnt());
    } finally {
        store.backdoor().getStatsCache().setRunInvalidatorEvery(5000);
        store.backdoor().getStatsCache().setMaxTimeInCache(500000);
        store.backdoor().getStatsCache().wakeInvalidator();
    }
}
Also used : ColumnStatistics(org.apache.hadoop.hive.metastore.api.ColumnStatistics) BooleanColumnStatsData(org.apache.hadoop.hive.metastore.api.BooleanColumnStatsData) Partition(org.apache.hadoop.hive.metastore.api.Partition) Table(org.apache.hadoop.hive.metastore.api.Table) AggrStats(org.apache.hadoop.hive.metastore.api.AggrStats) FieldSchema(org.apache.hadoop.hive.metastore.api.FieldSchema) SerDeInfo(org.apache.hadoop.hive.metastore.api.SerDeInfo) ArrayList(java.util.ArrayList) StorageDescriptor(org.apache.hadoop.hive.metastore.api.StorageDescriptor) ColumnStatisticsObj(org.apache.hadoop.hive.metastore.api.ColumnStatisticsObj) ColumnStatisticsDesc(org.apache.hadoop.hive.metastore.api.ColumnStatisticsDesc) ColumnStatisticsData(org.apache.hadoop.hive.metastore.api.ColumnStatisticsData) Test(org.junit.Test)

Example 63 with ColumnStatistics

use of org.apache.hadoop.hive.metastore.api.ColumnStatistics in project hive by apache.

the class TestHBaseStoreIntegration method tableStatistics.

@Test
public void tableStatistics() throws Exception {
    long now = System.currentTimeMillis();
    String dbname = "default";
    String tableName = "statstable";
    String boolcol = "boolcol";
    String longcol = "longcol";
    String doublecol = "doublecol";
    String stringcol = "stringcol";
    String binarycol = "bincol";
    String decimalcol = "deccol";
    long trues = 37;
    long falses = 12;
    long booleanNulls = 2;
    long longHigh = 120938479124L;
    long longLow = -12341243213412124L;
    long longNulls = 23;
    long longDVs = 213L;
    double doubleHigh = 123423.23423;
    double doubleLow = 0.00001234233;
    long doubleNulls = 92;
    long doubleDVs = 1234123421L;
    long strMaxLen = 1234;
    double strAvgLen = 32.3;
    long strNulls = 987;
    long strDVs = 906;
    long binMaxLen = 123412987L;
    double binAvgLen = 76.98;
    long binNulls = 976998797L;
    Decimal decHigh = new Decimal();
    decHigh.setScale((short) 3);
    // I have no clue how this is translated, but it
    decHigh.setUnscaled("3876".getBytes());
    // doesn't matter
    Decimal decLow = new Decimal();
    decLow.setScale((short) 3);
    decLow.setUnscaled("38".getBytes());
    long decNulls = 13;
    long decDVs = 923947293L;
    List<FieldSchema> cols = new ArrayList<FieldSchema>();
    cols.add(new FieldSchema(boolcol, "boolean", "nocomment"));
    cols.add(new FieldSchema(longcol, "long", "nocomment"));
    cols.add(new FieldSchema(doublecol, "double", "nocomment"));
    cols.add(new FieldSchema(stringcol, "varchar(32)", "nocomment"));
    cols.add(new FieldSchema(binarycol, "binary", "nocomment"));
    cols.add(new FieldSchema(decimalcol, "decimal(5, 3)", "nocomment"));
    SerDeInfo serde = new SerDeInfo("serde", "seriallib", null);
    StorageDescriptor sd = new StorageDescriptor(cols, "file:/tmp", "input", "output", false, 0, serde, null, null, emptyParameters);
    Table table = new Table(tableName, dbname, "me", (int) now / 1000, (int) now / 1000, 0, sd, null, emptyParameters, null, null, null);
    store.createTable(table);
    ColumnStatistics stats = new ColumnStatistics();
    ColumnStatisticsDesc desc = new ColumnStatisticsDesc();
    desc.setLastAnalyzed(now);
    desc.setDbName(dbname);
    desc.setTableName(tableName);
    desc.setIsTblLevel(true);
    stats.setStatsDesc(desc);
    // Do one column of each type
    ColumnStatisticsObj obj = new ColumnStatisticsObj();
    obj.setColName(boolcol);
    obj.setColType("boolean");
    ColumnStatisticsData data = new ColumnStatisticsData();
    BooleanColumnStatsData boolData = new BooleanColumnStatsData();
    boolData.setNumTrues(trues);
    boolData.setNumFalses(falses);
    boolData.setNumNulls(booleanNulls);
    data.setBooleanStats(boolData);
    obj.setStatsData(data);
    stats.addToStatsObj(obj);
    obj = new ColumnStatisticsObj();
    obj.setColName(longcol);
    obj.setColType("long");
    data = new ColumnStatisticsData();
    LongColumnStatsData longData = new LongColumnStatsData();
    longData.setHighValue(longHigh);
    longData.setLowValue(longLow);
    longData.setNumNulls(longNulls);
    longData.setNumDVs(longDVs);
    data.setLongStats(longData);
    obj.setStatsData(data);
    stats.addToStatsObj(obj);
    obj = new ColumnStatisticsObj();
    obj.setColName(doublecol);
    obj.setColType("double");
    data = new ColumnStatisticsData();
    DoubleColumnStatsData doubleData = new DoubleColumnStatsData();
    doubleData.setHighValue(doubleHigh);
    doubleData.setLowValue(doubleLow);
    doubleData.setNumNulls(doubleNulls);
    doubleData.setNumDVs(doubleDVs);
    data.setDoubleStats(doubleData);
    obj.setStatsData(data);
    stats.addToStatsObj(obj);
    store.updateTableColumnStatistics(stats);
    stats = store.getTableColumnStatistics(dbname, tableName, Arrays.asList(boolcol, longcol, doublecol));
    // We'll check all of the individual values later.
    Assert.assertEquals(3, stats.getStatsObjSize());
    // check that we can fetch just some of the columns
    stats = store.getTableColumnStatistics(dbname, tableName, Arrays.asList(boolcol));
    Assert.assertEquals(1, stats.getStatsObjSize());
    stats = new ColumnStatistics();
    stats.setStatsDesc(desc);
    obj = new ColumnStatisticsObj();
    obj.setColName(stringcol);
    obj.setColType("string");
    data = new ColumnStatisticsData();
    StringColumnStatsData strData = new StringColumnStatsData();
    strData.setMaxColLen(strMaxLen);
    strData.setAvgColLen(strAvgLen);
    strData.setNumNulls(strNulls);
    strData.setNumDVs(strDVs);
    data.setStringStats(strData);
    obj.setStatsData(data);
    stats.addToStatsObj(obj);
    obj = new ColumnStatisticsObj();
    obj.setColName(binarycol);
    obj.setColType("binary");
    data = new ColumnStatisticsData();
    BinaryColumnStatsData binData = new BinaryColumnStatsData();
    binData.setMaxColLen(binMaxLen);
    binData.setAvgColLen(binAvgLen);
    binData.setNumNulls(binNulls);
    data.setBinaryStats(binData);
    obj.setStatsData(data);
    stats.addToStatsObj(obj);
    obj = new ColumnStatisticsObj();
    obj.setColName(decimalcol);
    obj.setColType("decimal(5,3)");
    data = new ColumnStatisticsData();
    DecimalColumnStatsData decData = new DecimalColumnStatsData();
    LOG.debug("Setting decimal high value to " + decHigh.getScale() + " <" + new String(decHigh.getUnscaled()) + ">");
    decData.setHighValue(decHigh);
    decData.setLowValue(decLow);
    decData.setNumNulls(decNulls);
    decData.setNumDVs(decDVs);
    data.setDecimalStats(decData);
    obj.setStatsData(data);
    stats.addToStatsObj(obj);
    store.updateTableColumnStatistics(stats);
    stats = store.getTableColumnStatistics(dbname, tableName, Arrays.asList(boolcol, longcol, doublecol, stringcol, binarycol, decimalcol));
    Assert.assertEquals(now, stats.getStatsDesc().getLastAnalyzed());
    Assert.assertEquals(dbname, stats.getStatsDesc().getDbName());
    Assert.assertEquals(tableName, stats.getStatsDesc().getTableName());
    Assert.assertTrue(stats.getStatsDesc().isIsTblLevel());
    Assert.assertEquals(6, stats.getStatsObjSize());
    ColumnStatisticsData colData = stats.getStatsObj().get(0).getStatsData();
    Assert.assertEquals(ColumnStatisticsData._Fields.BOOLEAN_STATS, colData.getSetField());
    boolData = colData.getBooleanStats();
    Assert.assertEquals(trues, boolData.getNumTrues());
    Assert.assertEquals(falses, boolData.getNumFalses());
    Assert.assertEquals(booleanNulls, boolData.getNumNulls());
    colData = stats.getStatsObj().get(1).getStatsData();
    Assert.assertEquals(ColumnStatisticsData._Fields.LONG_STATS, colData.getSetField());
    longData = colData.getLongStats();
    Assert.assertEquals(longHigh, longData.getHighValue());
    Assert.assertEquals(longLow, longData.getLowValue());
    Assert.assertEquals(longNulls, longData.getNumNulls());
    Assert.assertEquals(longDVs, longData.getNumDVs());
    colData = stats.getStatsObj().get(2).getStatsData();
    Assert.assertEquals(ColumnStatisticsData._Fields.DOUBLE_STATS, colData.getSetField());
    doubleData = colData.getDoubleStats();
    Assert.assertEquals(doubleHigh, doubleData.getHighValue(), 0.01);
    Assert.assertEquals(doubleLow, doubleData.getLowValue(), 0.01);
    Assert.assertEquals(doubleNulls, doubleData.getNumNulls());
    Assert.assertEquals(doubleDVs, doubleData.getNumDVs());
    colData = stats.getStatsObj().get(3).getStatsData();
    Assert.assertEquals(ColumnStatisticsData._Fields.STRING_STATS, colData.getSetField());
    strData = colData.getStringStats();
    Assert.assertEquals(strMaxLen, strData.getMaxColLen());
    Assert.assertEquals(strAvgLen, strData.getAvgColLen(), 0.01);
    Assert.assertEquals(strNulls, strData.getNumNulls());
    Assert.assertEquals(strDVs, strData.getNumDVs());
    colData = stats.getStatsObj().get(4).getStatsData();
    Assert.assertEquals(ColumnStatisticsData._Fields.BINARY_STATS, colData.getSetField());
    binData = colData.getBinaryStats();
    Assert.assertEquals(binMaxLen, binData.getMaxColLen());
    Assert.assertEquals(binAvgLen, binData.getAvgColLen(), 0.01);
    Assert.assertEquals(binNulls, binData.getNumNulls());
    colData = stats.getStatsObj().get(5).getStatsData();
    Assert.assertEquals(ColumnStatisticsData._Fields.DECIMAL_STATS, colData.getSetField());
    decData = colData.getDecimalStats();
    Assert.assertEquals(decHigh, decData.getHighValue());
    Assert.assertEquals(decLow, decData.getLowValue());
    Assert.assertEquals(decNulls, decData.getNumNulls());
    Assert.assertEquals(decDVs, decData.getNumDVs());
}
Also used : ColumnStatistics(org.apache.hadoop.hive.metastore.api.ColumnStatistics) BooleanColumnStatsData(org.apache.hadoop.hive.metastore.api.BooleanColumnStatsData) Table(org.apache.hadoop.hive.metastore.api.Table) FieldSchema(org.apache.hadoop.hive.metastore.api.FieldSchema) SerDeInfo(org.apache.hadoop.hive.metastore.api.SerDeInfo) ArrayList(java.util.ArrayList) StorageDescriptor(org.apache.hadoop.hive.metastore.api.StorageDescriptor) StringColumnStatsData(org.apache.hadoop.hive.metastore.api.StringColumnStatsData) LongColumnStatsData(org.apache.hadoop.hive.metastore.api.LongColumnStatsData) BinaryColumnStatsData(org.apache.hadoop.hive.metastore.api.BinaryColumnStatsData) ColumnStatisticsObj(org.apache.hadoop.hive.metastore.api.ColumnStatisticsObj) DoubleColumnStatsData(org.apache.hadoop.hive.metastore.api.DoubleColumnStatsData) DecimalColumnStatsData(org.apache.hadoop.hive.metastore.api.DecimalColumnStatsData) Decimal(org.apache.hadoop.hive.metastore.api.Decimal) ColumnStatisticsDesc(org.apache.hadoop.hive.metastore.api.ColumnStatisticsDesc) ColumnStatisticsData(org.apache.hadoop.hive.metastore.api.ColumnStatisticsData) Test(org.junit.Test)

Example 64 with ColumnStatistics

use of org.apache.hadoop.hive.metastore.api.ColumnStatistics in project hive by apache.

the class TestHBaseSchemaTool method oneMondoTest.

@Test
public void oneMondoTest() throws Exception {
    // This is a pain to do in one big test, but we have to control the order so that we have tests
    // without dbs, etc.
    HBaseSchemaTool tool = new HBaseSchemaTool();
    ByteArrayOutputStream outStr = new ByteArrayOutputStream();
    PrintStream out = new PrintStream(outStr);
    ByteArrayOutputStream errStr = new ByteArrayOutputStream();
    PrintStream err = new PrintStream(errStr);
    // This needs to be up front before we create any tables or partitions
    tool.go(false, HBaseReadWrite.SD_TABLE, null, "whatever", conf, out, err);
    Assert.assertEquals("No storage descriptors" + lsep, outStr.toString());
    // This one needs to be up front too
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.SEQUENCES_TABLE, null, "whatever", conf, out, err);
    Assert.assertEquals("No sequences" + lsep, outStr.toString());
    // Create some databases
    String[] dbNames = new String[3];
    for (int i = 0; i < dbNames.length; i++) {
        dbNames[i] = "db" + i;
        Database db = new Database(dbNames[i], "no description", "file:///tmp", emptyParameters);
        store.createDatabase(db);
    }
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.DB_TABLE, "db0", null, conf, out, err);
    Assert.assertEquals("{\"name\":\"db0\",\"description\":\"no description\"," + "\"locationUri\":\"file:///tmp\",\"parameters\":{}}" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.DB_TABLE, null, ".*", conf, out, err);
    Assert.assertEquals("{\"name\":\"db0\",\"description\":\"no description\"," + "\"locationUri\":\"file:///tmp\",\"parameters\":{}}" + lsep + "{\"name\":\"db1\",\"description\":\"no description\"," + "\"locationUri\":\"file:///tmp\",\"parameters\":{}}" + lsep + "{\"name\":\"db2\",\"description\":\"no description\"," + "\"locationUri\":\"file:///tmp\",\"parameters\":{}}" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.DB_TABLE, null, "db[12]", conf, out, err);
    Assert.assertEquals("{\"name\":\"db1\",\"description\":\"no description\"," + "\"locationUri\":\"file:///tmp\",\"parameters\":{}}" + lsep + "{\"name\":\"db2\",\"description\":\"no description\"," + "\"locationUri\":\"file:///tmp\",\"parameters\":{}}" + lsep, outStr.toString());
    String[] roleNames = new String[2];
    for (int i = 0; i < roleNames.length; i++) {
        roleNames[i] = "role" + i;
        store.addRole(roleNames[i], "me");
    }
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.ROLE_TABLE, null, "role.", conf, out, err);
    Assert.assertEquals("{\"roleName\":\"role0\",\"createTime\":now,\"ownerName\":\"me\"}" + lsep + "{\"roleName\":\"role1\",\"createTime\":now,\"ownerName\":\"me\"}" + lsep, outStr.toString().replaceAll("createTime\":[0-9]+", "createTime\":now"));
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.ROLE_TABLE, "role1", null, conf, out, err);
    Assert.assertEquals("{\"roleName\":\"role1\",\"createTime\":now,\"ownerName\":\"me\"}" + lsep, outStr.toString().replaceAll("createTime\":[0-9]+", "createTime\":now"));
    Role role1 = store.getRole("role1");
    store.grantRole(role1, "fred", PrincipalType.USER, "me", PrincipalType.USER, false);
    store.grantRole(role1, "joanne", PrincipalType.USER, "me", PrincipalType.USER, false);
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.USER_TO_ROLE_TABLE, null, ".*", conf, out, err);
    Assert.assertEquals("fred: role1" + lsep + "joanne: role1" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.USER_TO_ROLE_TABLE, "joanne", null, conf, out, err);
    Assert.assertEquals("role1" + lsep, outStr.toString());
    String[] funcNames = new String[3];
    for (int i = 0; i < funcNames.length; i++) {
        funcNames[i] = "func" + i;
        Function function = new Function(funcNames[i], "db1", "Function", "me", PrincipalType.USER, 0, FunctionType.JAVA, null);
        store.createFunction(function);
    }
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.FUNC_TABLE, "db1.func0", null, conf, out, err);
    Assert.assertEquals("{\"functionName\":\"func0\",\"dbName\":\"db1\"," + "\"className\":\"Function\",\"ownerName\":\"me\",\"ownerType\":1,\"createTime\":0," + "\"functionType\":1}" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.FUNC_TABLE, null, ".*", conf, out, err);
    Assert.assertEquals("{\"functionName\":\"func0\",\"dbName\":\"db1\"," + "\"className\":\"Function\",\"ownerName\":\"me\",\"ownerType\":1,\"createTime\":0," + "\"functionType\":1}" + lsep + "{\"functionName\":\"func1\",\"dbName\":\"db1\"," + "\"className\":\"Function\",\"ownerName\":\"me\",\"ownerType\":1,\"createTime\":0," + "\"functionType\":1}" + lsep + "{\"functionName\":\"func2\",\"dbName\":\"db1\"," + "\"className\":\"Function\",\"ownerName\":\"me\",\"ownerType\":1,\"createTime\":0," + "\"functionType\":1}" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.FUNC_TABLE, null, "db1.func[12]", conf, out, err);
    Assert.assertEquals("{\"functionName\":\"func1\",\"dbName\":\"db1\"," + "\"className\":\"Function\",\"ownerName\":\"me\",\"ownerType\":1,\"createTime\":0," + "\"functionType\":1}" + lsep + "{\"functionName\":\"func2\",\"dbName\":\"db1\"," + "\"className\":\"Function\",\"ownerName\":\"me\",\"ownerType\":1,\"createTime\":0," + "\"functionType\":1}" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.GLOBAL_PRIVS_TABLE, null, null, conf, out, err);
    Assert.assertEquals("No global privileges" + lsep, outStr.toString());
    List<HiveObjectPrivilege> privileges = new ArrayList<>();
    HiveObjectRef hiveObjRef = new HiveObjectRef(HiveObjectType.GLOBAL, "db0", "tab0", null, null);
    PrivilegeGrantInfo grantInfo = new PrivilegeGrantInfo("read", 0, "me", PrincipalType.USER, false);
    HiveObjectPrivilege hop = new HiveObjectPrivilege(hiveObjRef, "user", PrincipalType.USER, grantInfo);
    privileges.add(hop);
    grantInfo = new PrivilegeGrantInfo("create", 0, "me", PrincipalType.USER, true);
    hop = new HiveObjectPrivilege(hiveObjRef, "user", PrincipalType.USER, grantInfo);
    privileges.add(hop);
    PrivilegeBag pBag = new PrivilegeBag(privileges);
    store.grantPrivileges(pBag);
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.GLOBAL_PRIVS_TABLE, null, null, conf, out, err);
    Assert.assertEquals("{\"userPrivileges\":{\"user\":[{\"privilege\":\"read\",\"createTime\":0," + "\"grantor\":\"me\",\"grantorType\":1,\"grantOption\":0},{\"privilege\":\"create\"," + "\"createTime\":0,\"grantor\":\"me\",\"grantorType\":1,\"grantOption\":1}]}}" + lsep, outStr.toString());
    String[] tableNames = new String[3];
    for (int i = 0; i < tableNames.length; i++) {
        tableNames[i] = "tab" + i;
        StorageDescriptor sd = new StorageDescriptor(Arrays.asList(new FieldSchema("col1", "int", ""), new FieldSchema("col2", "varchar(32)", "")), "/tmp", null, null, false, 0, null, null, null, Collections.<String, String>emptyMap());
        Table tab = new Table(tableNames[i], dbNames[0], "me", 0, 0, 0, sd, Arrays.asList(new FieldSchema("pcol1", "string", ""), new FieldSchema("pcol2", "string", "")), Collections.<String, String>emptyMap(), null, null, null);
        store.createTable(tab);
    }
    ColumnStatisticsDesc tableStatsDesc = new ColumnStatisticsDesc(false, "db0", "tab0");
    ColumnStatisticsData tcsd = new ColumnStatisticsData();
    LongColumnStatsData tlcsd = new LongColumnStatsData(1, 2);
    tlcsd.setLowValue(-95);
    tlcsd.setHighValue(95);
    tcsd.setLongStats(tlcsd);
    ColumnStatisticsData tcsd2 = new ColumnStatisticsData();
    tcsd2.setStringStats(new StringColumnStatsData(97, 18.78, 29, 397));
    List<ColumnStatisticsObj> tcsos = Arrays.asList(new ColumnStatisticsObj("col1", "int", tcsd), new ColumnStatisticsObj("col2", "varchar(32)", tcsd2));
    ColumnStatistics tStatObj = new ColumnStatistics(tableStatsDesc, tcsos);
    store.updateTableColumnStatistics(tStatObj);
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.TABLE_TABLE, "db0.tab1", null, conf, out, err);
    Assert.assertEquals("{\"tableName\":\"tab1\",\"dbName\":\"db0\",\"owner\":\"me\"," + "\"createTime\":0,\"lastAccessTime\":0,\"retention\":0," + "\"partitionKeys\":[{\"name\":\"pcol1\",\"type\":\"string\",\"comment\":\"\"}," + "{\"name\":\"pcol2\",\"type\":\"string\",\"comment\":\"\"}],\"parameters\":{}," + "\"tableType\":\"\",\"rewriteEnabled\":0} sdHash: qQTgZAi5VzgpozzFGmIVTQ stats:" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.TABLE_TABLE, null, "db0.*", conf, out, err);
    Assert.assertEquals("{\"tableName\":\"tab0\",\"dbName\":\"db0\",\"owner\":\"me\"," + "\"createTime\":0,\"lastAccessTime\":0,\"retention\":0," + "\"partitionKeys\":[{\"name\":\"pcol1\",\"type\":\"string\",\"comment\":\"\"}," + "{\"name\":\"pcol2\",\"type\":\"string\",\"comment\":\"\"}],\"parameters\":{\"COLUMN_STATS_ACCURATE\":\"{\\\"COLUMN_STATS\\\":{\\\"col1\\\":\\\"true\\\",\\\"col2\\\":\\\"true\\\"}}\"}," + "\"tableType\":\"\",\"rewriteEnabled\":0} sdHash: qQTgZAi5VzgpozzFGmIVTQ stats: column " + "col1: {\"colName\":\"col1\",\"colType\":\"int\"," + "\"statsData\":{\"longStats\":{\"lowValue\":-95,\"highValue\":95,\"numNulls\":1," + "\"numDVs\":2,\"bitVectors\":\"\"}}} column col2: {\"colName\":\"col2\",\"colType\":\"varchar(32)\"," + "\"statsData\":{\"stringStats\":{\"maxColLen\":97,\"avgColLen\":18.78," + "\"numNulls\":29,\"numDVs\":397,\"bitVectors\":\"\"}}}" + lsep + "{\"tableName\":\"tab1\",\"dbName\":\"db0\",\"owner\":\"me\",\"createTime\":0," + "\"lastAccessTime\":0,\"retention\":0,\"partitionKeys\":[{\"name\":\"pcol1\"," + "\"type\":\"string\",\"comment\":\"\"},{\"name\":\"pcol2\",\"type\":\"string\"," + "\"comment\":\"\"}],\"parameters\":{},\"tableType\":\"\",\"rewriteEnabled\":0} sdHash: " + "qQTgZAi5VzgpozzFGmIVTQ stats:" + lsep + "{\"tableName\":\"tab2\",\"dbName\":\"db0\",\"owner\":\"me\",\"createTime\":0," + "\"lastAccessTime\":0,\"retention\":0,\"partitionKeys\":[{\"name\":\"pcol1\"," + "\"type\":\"string\",\"comment\":\"\"},{\"name\":\"pcol2\",\"type\":\"string\"," + "\"comment\":\"\"}],\"parameters\":{},\"tableType\":\"\",\"rewriteEnabled\":0} sdHash: " + "qQTgZAi5VzgpozzFGmIVTQ stats:" + lsep, outStr.toString());
    List<List<String>> partVals = Arrays.asList(Arrays.asList("a", "b"), Arrays.asList("c", "d"));
    for (List<String> pv : partVals) {
        StorageDescriptor sd = new StorageDescriptor(Arrays.asList(new FieldSchema("col1", "int", ""), new FieldSchema("col2", "varchar(32)", "")), "/tmp", null, null, false, 0, null, null, null, Collections.<String, String>emptyMap());
        Partition p = new Partition(pv, "db0", "tab1", 0, 0, sd, Collections.<String, String>emptyMap());
        store.addPartition(p);
    }
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.PART_TABLE, "db0.tab1.a.b", null, conf, out, err);
    Assert.assertEquals("{\"values\":[\"a\",\"b\"],\"dbName\":\"db0\",\"tableName\":\"tab1\"," + "\"createTime\":0,\"lastAccessTime\":0,\"parameters\":{}} sdHash: " + "qQTgZAi5VzgpozzFGmIVTQ stats:" + lsep, outStr.toString());
    ColumnStatisticsDesc statsDesc = new ColumnStatisticsDesc(false, "db0", "tab1");
    statsDesc.setPartName("pcol1=c/pcol2=d");
    ColumnStatisticsData csd1 = new ColumnStatisticsData();
    LongColumnStatsData lcsd = new LongColumnStatsData(1, 2);
    lcsd.setLowValue(-95);
    lcsd.setHighValue(95);
    csd1.setLongStats(lcsd);
    ColumnStatisticsData csd2 = new ColumnStatisticsData();
    csd2.setStringStats(new StringColumnStatsData(97, 18.78, 29, 397));
    List<ColumnStatisticsObj> csos = Arrays.asList(new ColumnStatisticsObj("col1", "int", csd1), new ColumnStatisticsObj("col2", "varchar(32)", csd2));
    ColumnStatistics statsObj = new ColumnStatistics(statsDesc, csos);
    store.updatePartitionColumnStatistics(statsObj, partVals.get(1));
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.PART_TABLE, "db0.tab1.c.d", null, conf, out, err);
    Assert.assertEquals("{\"values\":[\"c\",\"d\"],\"dbName\":\"db0\",\"tableName\":\"tab1\"," + "\"createTime\":0,\"lastAccessTime\":0,\"parameters\":{\"COLUMN_STATS_ACCURATE\":\"{\\\"COLUMN_STATS\\\":{\\\"col1\\\":\\\"true\\\",\\\"col2\\\":\\\"true\\\"}}\"}} sdHash: qQTgZAi5VzgpozzFGmIVTQ " + "stats: column col1: {\"colName\":\"col1\",\"colType\":\"int\"," + "\"statsData\":{\"longStats\":{\"lowValue\":-95,\"highValue\":95,\"numNulls\":1," + "\"numDVs\":2,\"bitVectors\":\"\"}}} column col2: {\"colName\":\"col2\",\"colType\":\"varchar(32)\"," + "\"statsData\":{\"stringStats\":{\"maxColLen\":97,\"avgColLen\":18.78,\"numNulls\":29," + "\"numDVs\":397,\"bitVectors\":\"\"}}}" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.PART_TABLE, null, "db0.tab1", conf, out, err);
    Assert.assertEquals("{\"values\":[\"a\",\"b\"],\"dbName\":\"db0\",\"tableName\":\"tab1\"," + "\"createTime\":0,\"lastAccessTime\":0,\"parameters\":{}} sdHash: qQTgZAi5VzgpozzFGmIVTQ " + "stats:" + lsep + "{\"values\":[\"c\",\"d\"],\"dbName\":\"db0\",\"tableName\":\"tab1\",\"createTime\":0," + "\"lastAccessTime\":0,\"parameters\":{\"COLUMN_STATS_ACCURATE\":\"{\\\"COLUMN_STATS\\\":{\\\"col1\\\":\\\"true\\\",\\\"col2\\\":\\\"true\\\"}}\"}} sdHash: qQTgZAi5VzgpozzFGmIVTQ stats: column " + "col1: {\"colName\":\"col1\",\"colType\":\"int\"," + "\"statsData\":{\"longStats\":{\"lowValue\":-95,\"highValue\":95,\"numNulls\":1," + "\"numDVs\":2,\"bitVectors\":\"\"}}} column col2: {\"colName\":\"col2\",\"colType\":\"varchar(32)\"," + "\"statsData\":{\"stringStats\":{\"maxColLen\":97,\"avgColLen\":18.78,\"numNulls\":29," + "\"numDVs\":397,\"bitVectors\":\"\"}}}" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.PART_TABLE, null, "db0.tab1.a", conf, out, err);
    Assert.assertEquals("{\"values\":[\"a\",\"b\"],\"dbName\":\"db0\",\"tableName\":\"tab1\"," + "\"createTime\":0,\"lastAccessTime\":0,\"parameters\":{}} sdHash: qQTgZAi5VzgpozzFGmIVTQ " + "stats:" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.SD_TABLE, "qQTgZAi5VzgpozzFGmIVTQ", null, conf, out, err);
    Assert.assertEquals("{\"cols\":[{\"name\":\"col1\",\"type\":\"int\",\"comment\":\"\"}," + "{\"name\":\"col2\",\"type\":\"varchar(32)\",\"comment\":\"\"}],\"compressed\":0," + "\"numBuckets\":0,\"bucketCols\":[],\"sortCols\":[],\"storedAsSubDirectories\":0}" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.SD_TABLE, null, "whatever", conf, out, err);
    Assert.assertEquals("qQTgZAi5VzgpozzFGmIVTQ: {\"cols\":[{\"name\":\"col1\",\"type\":\"int\"," + "\"comment\":\"\"}," + "{\"name\":\"col2\",\"type\":\"varchar(32)\",\"comment\":\"\"}],\"compressed\":0," + "\"numBuckets\":0,\"bucketCols\":[],\"sortCols\":[],\"storedAsSubDirectories\":0}" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.SECURITY_TABLE, null, "whatever", conf, out, err);
    Assert.assertEquals("No security related entries" + lsep, outStr.toString());
    store.addMasterKey("this be a key");
    store.addToken("tokenid", "delegation token");
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.SECURITY_TABLE, null, "whatever", conf, out, err);
    Assert.assertEquals("Master key 0: this be a key" + lsep + "Delegation token tokenid: delegation token" + lsep, outStr.toString());
    outStr = new ByteArrayOutputStream();
    out = new PrintStream(outStr);
    tool.go(false, HBaseReadWrite.SEQUENCES_TABLE, null, "whatever", conf, out, err);
    Assert.assertEquals("master_key: 1" + lsep, outStr.toString());
}
Also used : PrivilegeBag(org.apache.hadoop.hive.metastore.api.PrivilegeBag) FieldSchema(org.apache.hadoop.hive.metastore.api.FieldSchema) ArrayList(java.util.ArrayList) StorageDescriptor(org.apache.hadoop.hive.metastore.api.StorageDescriptor) Function(org.apache.hadoop.hive.metastore.api.Function) ColumnStatisticsObj(org.apache.hadoop.hive.metastore.api.ColumnStatisticsObj) ColumnStatisticsDesc(org.apache.hadoop.hive.metastore.api.ColumnStatisticsDesc) Database(org.apache.hadoop.hive.metastore.api.Database) ArrayList(java.util.ArrayList) List(java.util.List) ColumnStatistics(org.apache.hadoop.hive.metastore.api.ColumnStatistics) PrintStream(java.io.PrintStream) Partition(org.apache.hadoop.hive.metastore.api.Partition) Table(org.apache.hadoop.hive.metastore.api.Table) PrivilegeGrantInfo(org.apache.hadoop.hive.metastore.api.PrivilegeGrantInfo) HiveObjectRef(org.apache.hadoop.hive.metastore.api.HiveObjectRef) StringColumnStatsData(org.apache.hadoop.hive.metastore.api.StringColumnStatsData) ByteArrayOutputStream(java.io.ByteArrayOutputStream) LongColumnStatsData(org.apache.hadoop.hive.metastore.api.LongColumnStatsData) Role(org.apache.hadoop.hive.metastore.api.Role) HiveObjectPrivilege(org.apache.hadoop.hive.metastore.api.HiveObjectPrivilege) ColumnStatisticsData(org.apache.hadoop.hive.metastore.api.ColumnStatisticsData) Test(org.junit.Test)

Example 65 with ColumnStatistics

use of org.apache.hadoop.hive.metastore.api.ColumnStatistics in project hive by apache.

the class TestHBaseAggregateStatsNDVUniformDist method TwoEndsOfPartitionsHaveBitVectorStatus.

@Test
public void TwoEndsOfPartitionsHaveBitVectorStatus() throws Exception {
    String dbName = "default";
    String tableName = "snp";
    long now = System.currentTimeMillis();
    List<FieldSchema> cols = new ArrayList<>();
    cols.add(new FieldSchema("col3", "long", "nocomment"));
    SerDeInfo serde = new SerDeInfo("serde", "seriallib", null);
    StorageDescriptor sd = new StorageDescriptor(cols, "file:/tmp", "input", "output", false, 0, serde, null, null, Collections.<String, String>emptyMap());
    List<FieldSchema> partCols = new ArrayList<>();
    partCols.add(new FieldSchema("ds", "string", ""));
    Table table = new Table(tableName, dbName, "me", (int) now, (int) now, 0, sd, partCols, Collections.<String, String>emptyMap(), null, null, null);
    store.createTable(table);
    List<List<String>> partVals = new ArrayList<>();
    for (int i = 0; i < 10; i++) {
        List<String> partVal = Arrays.asList("" + i);
        partVals.add(partVal);
        StorageDescriptor psd = new StorageDescriptor(sd);
        psd.setLocation("file:/tmp/default/hit/ds=" + partVal);
        Partition part = new Partition(partVal, dbName, tableName, (int) now, (int) now, psd, Collections.<String, String>emptyMap());
        store.addPartition(part);
        if (i < 2 || i > 7) {
            ColumnStatistics cs = new ColumnStatistics();
            ColumnStatisticsDesc desc = new ColumnStatisticsDesc(false, dbName, tableName);
            desc.setLastAnalyzed(now);
            desc.setPartName("ds=" + partVal);
            cs.setStatsDesc(desc);
            ColumnStatisticsObj obj = new ColumnStatisticsObj();
            obj.setColName("col3");
            obj.setColType("long");
            ColumnStatisticsData data = new ColumnStatisticsData();
            LongColumnStatsData dcsd = new LongColumnStatsData();
            dcsd.setHighValue(1000 + i);
            dcsd.setLowValue(-1000 - i);
            dcsd.setNumNulls(i);
            dcsd.setNumDVs(10 * i + 1);
            dcsd.setBitVectors(bitVectors[i / 5]);
            data.setLongStats(dcsd);
            obj.setStatsData(data);
            cs.addToStatsObj(obj);
            store.updatePartitionColumnStatistics(cs, partVal);
        }
    }
    Checker statChecker = new Checker() {

        @Override
        public void checkStats(AggrStats aggrStats) throws Exception {
            Assert.assertEquals(4, aggrStats.getPartsFound());
            Assert.assertEquals(1, aggrStats.getColStatsSize());
            ColumnStatisticsObj cso = aggrStats.getColStats().get(0);
            Assert.assertEquals("col3", cso.getColName());
            Assert.assertEquals("long", cso.getColType());
            LongColumnStatsData lcsd = cso.getStatsData().getLongStats();
            Assert.assertEquals(1010, lcsd.getHighValue(), 0.01);
            Assert.assertEquals(-1010, lcsd.getLowValue(), 0.01);
            Assert.assertEquals(45, lcsd.getNumNulls());
            Assert.assertEquals(12, lcsd.getNumDVs());
        }
    };
    List<String> partNames = new ArrayList<>();
    for (int i = 0; i < 10; i++) {
        partNames.add("ds=" + i);
    }
    AggrStats aggrStats = store.get_aggr_stats_for(dbName, tableName, partNames, Arrays.asList("col3"));
    statChecker.checkStats(aggrStats);
}
Also used : ColumnStatistics(org.apache.hadoop.hive.metastore.api.ColumnStatistics) Partition(org.apache.hadoop.hive.metastore.api.Partition) Table(org.apache.hadoop.hive.metastore.api.Table) AggrStats(org.apache.hadoop.hive.metastore.api.AggrStats) FieldSchema(org.apache.hadoop.hive.metastore.api.FieldSchema) SerDeInfo(org.apache.hadoop.hive.metastore.api.SerDeInfo) ArrayList(java.util.ArrayList) StorageDescriptor(org.apache.hadoop.hive.metastore.api.StorageDescriptor) LongColumnStatsData(org.apache.hadoop.hive.metastore.api.LongColumnStatsData) ColumnStatisticsObj(org.apache.hadoop.hive.metastore.api.ColumnStatisticsObj) ColumnStatisticsDesc(org.apache.hadoop.hive.metastore.api.ColumnStatisticsDesc) ArrayList(java.util.ArrayList) List(java.util.List) ColumnStatisticsData(org.apache.hadoop.hive.metastore.api.ColumnStatisticsData) Test(org.junit.Test)

Aggregations

ColumnStatistics (org.apache.hadoop.hive.metastore.api.ColumnStatistics)90 ColumnStatisticsObj (org.apache.hadoop.hive.metastore.api.ColumnStatisticsObj)75 ColumnStatisticsDesc (org.apache.hadoop.hive.metastore.api.ColumnStatisticsDesc)67 ColumnStatisticsData (org.apache.hadoop.hive.metastore.api.ColumnStatisticsData)62 ArrayList (java.util.ArrayList)61 Test (org.junit.Test)53 FieldSchema (org.apache.hadoop.hive.metastore.api.FieldSchema)40 Table (org.apache.hadoop.hive.metastore.api.Table)38 Partition (org.apache.hadoop.hive.metastore.api.Partition)33 StorageDescriptor (org.apache.hadoop.hive.metastore.api.StorageDescriptor)32 SerDeInfo (org.apache.hadoop.hive.metastore.api.SerDeInfo)31 AggrStats (org.apache.hadoop.hive.metastore.api.AggrStats)29 List (java.util.List)26 LongColumnStatsData (org.apache.hadoop.hive.metastore.api.LongColumnStatsData)19 MetaException (org.apache.hadoop.hive.metastore.api.MetaException)14 BooleanColumnStatsData (org.apache.hadoop.hive.metastore.api.BooleanColumnStatsData)13 DoubleColumnStatsData (org.apache.hadoop.hive.metastore.api.DoubleColumnStatsData)12 StringColumnStatsData (org.apache.hadoop.hive.metastore.api.StringColumnStatsData)12 HashMap (java.util.HashMap)11 Database (org.apache.hadoop.hive.metastore.api.Database)9