Examples with VectorizedParquetRecordReader org.apache.hadoop.hive.ql.io.parquet.vector.VectorizedParquetRecordReader used on opensource projects

Example 26 with VectorizedParquetRecordReader

use of org.apache.hadoop.hive.ql.io.parquet.vector.VectorizedParquetRecordReader in project hive by apache.

the class VectorizedColumnReaderTestBase method createTestParquetReader.

public static VectorizedParquetRecordReader createTestParquetReader(String schemaString, Configuration conf) throws IOException, InterruptedException, HiveException {
    conf.set(PARQUET_READ_SCHEMA, schemaString);
    HiveConf.setBoolVar(conf, HiveConf.ConfVars.HIVE_VECTORIZATION_ENABLED, true);
    HiveConf.setVar(conf, HiveConf.ConfVars.PLAN, "//tmp");
    Job vectorJob = new Job(conf, "read vector");
    ParquetInputFormat.setInputPaths(vectorJob, file);
    initialVectorizedRowBatchCtx(conf);
    return new VectorizedParquetRecordReader(getFileSplit(vectorJob), new JobConf(conf));
}

Example 27 with VectorizedParquetRecordReader

use of org.apache.hadoop.hive.ql.io.parquet.vector.VectorizedParquetRecordReader in project hive by apache.

the class VectorizedColumnReaderTestBase method booleanRead.

protected void booleanRead() throws Exception {
    Configuration conf = new Configuration();
    conf.set(IOConstants.COLUMNS, "boolean_field");
    conf.set(IOConstants.COLUMNS_TYPES, "boolean");
    conf.setBoolean(ColumnProjectionUtils.READ_ALL_COLUMNS, false);
    conf.set(ColumnProjectionUtils.READ_COLUMN_IDS_CONF_STR, "0");
    VectorizedParquetRecordReader reader = createTestParquetReader("message test { required boolean boolean_field;}", conf);
    VectorizedRowBatch previous = reader.createValue();
    try {
        int c = 0;
        while (reader.next(NullWritable.get(), previous)) {
            LongColumnVector vector = (LongColumnVector) previous.cols[0];
            assertTrue(vector.noNulls);
            for (int i = 0; i < vector.vector.length; i++) {
                if (c == nElements) {
                    break;
                }
                assertEquals("Failed at " + c, (getBooleanValue(c) ? 1 : 0), vector.vector[i]);
                assertFalse(vector.isNull[i]);
                c++;
            }
        }
        assertEquals(nElements, c);
    } finally {
        reader.close();
    }
}

Example 28 with VectorizedParquetRecordReader

use of org.apache.hadoop.hive.ql.io.parquet.vector.VectorizedParquetRecordReader in project hive by apache.

the class VectorizedColumnReaderTestBase method floatReadInt.

protected void floatReadInt(boolean isDictionaryEncoding) throws InterruptedException, HiveException, IOException {
    conf.set(IOConstants.COLUMNS, "int32_field");
    conf.set(IOConstants.COLUMNS_TYPES, "float");
    conf.setBoolean(ColumnProjectionUtils.READ_ALL_COLUMNS, false);
    conf.set(ColumnProjectionUtils.READ_COLUMN_IDS_CONF_STR, "0");
    VectorizedParquetRecordReader reader = createTestParquetReader("message test { required int32" + " int32_field;}", conf);
    VectorizedRowBatch previous = reader.createValue();
    try {
        int c = 0;
        while (reader.next(NullWritable.get(), previous)) {
            DoubleColumnVector vector = (DoubleColumnVector) previous.cols[0];
            assertTrue(vector.noNulls);
            for (int i = 0; i < vector.vector.length; i++) {
                if (c == nElements) {
                    break;
                }
                assertEquals("Failed at " + c, getIntValue(isDictionaryEncoding, c), vector.vector[i], 0);
                assertFalse(vector.isNull[i]);
                c++;
            }
        }
        assertEquals(nElements, c);
    } finally {
        reader.close();
    }
}