Examples with Table com.google.cloud.teleport.v2.templates.spanner.ddl.Table used on opensource projects

Example 46 with Table

use of com.google.cloud.teleport.v2.templates.spanner.ddl.Table in project DataflowTemplates by GoogleCloudPlatform.

the class DataplexBigQueryToGcsFilterTest method test_whenBeforeDateSet_filterExcludesTablesAndPartitions.

@Test
public void test_whenBeforeDateSet_filterExcludesTablesAndPartitions() {
    BigQueryTable.Builder olderTable = table().setLastModificationTime(TS_MICROS_2021_01_01_15_00_00_UTC - 1000L);
    BigQueryTable.Builder newerTable = table().setLastModificationTime(TS_MICROS_2021_01_01_15_00_00_UTC + 1000L);
    BigQueryTablePartition olderPartition = partition().setPartitionName("p1").setLastModificationTime(TS_MICROS_2021_01_01_15_00_00_UTC - 1000L).build();
    BigQueryTablePartition newerPartition = partition().setPartitionName("p2").setLastModificationTime(TS_MICROS_2021_01_01_15_00_00_UTC + 1000L).build();
    List<BigQueryTablePartition> partitions = Arrays.asList(olderPartition, newerPartition);
    options.setTables(null);
    options.setExportDataModifiedBeforeDateTime("2021-01-01T15:00:00Z");
    Filter f = new DataplexBigQueryToGcsFilter(options, new ArrayList<String>());
    assertThat(f.shouldSkipUnpartitionedTable(newerTable)).isTrue();
    assertThat(f.shouldSkipUnpartitionedTable(olderTable)).isFalse();
    // If a table is partitioned, we should filter individual partitions by modification time,
    // so the table itself should NOT be skipped no matter what the table modification time is.
    // Expecting shouldSkip = false for both newer and older tables:
    assertThat(f.shouldSkipPartitionedTable(newerTable, partitions)).isFalse();
    assertThat(f.shouldSkipPartitionedTable(olderTable, partitions)).isFalse();
    assertThat(f.shouldSkipPartition(olderTable, newerPartition)).isTrue();
    assertThat(f.shouldSkipPartition(olderTable, olderPartition)).isFalse();
}

Example 47 with Table

use of com.google.cloud.teleport.v2.templates.spanner.ddl.Table in project DataflowTemplates by GoogleCloudPlatform.

the class DataplexBigQueryToGcsFilterTest method test_whenTargetFileExistsWithWriteDispositionSKIP_filterExcludesTables.

@Test
public void test_whenTargetFileExistsWithWriteDispositionSKIP_filterExcludesTables() {
    BigQueryTable.Builder t = table().setTableName("table1").setPartitioningColumn("p2");
    BigQueryTablePartition p = partition().setPartitionName("partition1").build();
    options.setTables(null);
    options.setExportDataModifiedBeforeDateTime(null);
    options.setFileFormat(FileFormatOptions.AVRO);
    options.setWriteDisposition(WriteDispositionOptions.SKIP);
    Filter f = new DataplexBigQueryToGcsFilter(options, Arrays.asList("table1/output-table1.avro", "table1/p2=partition1/output-table1-partition1.avro"));
    assertThat(f.shouldSkipUnpartitionedTable(t)).isTrue();
    assertThat(f.shouldSkipPartition(t, p)).isTrue();
}

Example 48 with Table

use of com.google.cloud.teleport.v2.templates.spanner.ddl.Table in project DataflowTemplates by GoogleCloudPlatform.

the class DataplexBigQueryToGcsFilterTest method test_whenBeforeDateIs1DayDuration_dateParsedCorrectly.

@Test
public void test_whenBeforeDateIs1DayDuration_dateParsedCorrectly() {
    // current time in the DEFAULT time zone minus one day:
    long micros = Instant.now().minus(Duration.standardDays(1)).getMillis() * 1000L;
    BigQueryTable.Builder olderTable = table().setLastModificationTime(micros - 100000L);
    BigQueryTable.Builder newerTable = table().setLastModificationTime(micros + 100000L);
    options.setTables(null);
    options.setExportDataModifiedBeforeDateTime("-P1D");
    Filter f = new DataplexBigQueryToGcsFilter(options, new ArrayList<String>());
    assertThat(f.shouldSkipUnpartitionedTable(newerTable)).isTrue();
    assertThat(f.shouldSkipUnpartitionedTable(olderTable)).isFalse();
}

Example 49 with Table

use of com.google.cloud.teleport.v2.templates.spanner.ddl.Table in project DataflowTemplates by GoogleCloudPlatform.

the class BigQueryMetadataLoader method loadTablePartitions.

private List<BigQueryTablePartition> loadTablePartitions(BigQueryTable.Builder table, Filter filter) throws InterruptedException {
    String partitionSql = String.format("select partition_id, last_modified_time\n" + "from `%s.%s.INFORMATION_SCHEMA.PARTITIONS`\n" + "where table_name = @table_name", table.getProject(), table.getDataset());
    TableResult partitionRows = bqClient.query(QueryJobConfiguration.newBuilder(partitionSql).addNamedParameter("table_name", QueryParameterValue.string(table.getTableName())).build());
    List<BigQueryTablePartition> partitions = new ArrayList<>();
    partitionRows.iterateAll().forEach(// TODO(an2x): Check we didn't get duplicate partition names.
    r -> {
        BigQueryTablePartition p = BigQueryTablePartition.builder().setPartitionName(r.get(0).getStringValue()).setLastModificationTime(r.get(1).getTimestampValue()).build();
        if (filter == null || !filter.shouldSkipPartition(table, p)) {
            partitions.add(p);
        }
    });
    return partitions;
}

Example 50 with Table

use of com.google.cloud.teleport.v2.templates.spanner.ddl.Table in project DataflowTemplates by GoogleCloudPlatform.

the class BigQueryMetadataLoader method loadDatasetMetadata.

/**
 * Loads metadata for all tables in the dataset {@code datasetId} returning only those accepted by
 * the {@code filter}.
 *
 * @param filter if {@code null}, will include all tables and partitions
 */
public List<BigQueryTable> loadDatasetMetadata(DatasetId datasetId, @Nullable Filter filter) throws InterruptedException, ExecutionException {
    String tableSql = String.format("select\n" + "    table_id,\n" + "    timestamp_millis(last_modified_time) as last_modified_time,\n" + "    (select column_name from `%s.%s.INFORMATION_SCHEMA.COLUMNS` c\n" + "      where c.table_catalog = t.project_id\n" + "        and c.table_schema = t.dataset_id\n" + "        and c.table_name = t.table_id\n" + "        and c.is_partitioning_column = 'YES') as partitioning_column,\n" + "  from `%s.%s.__TABLES__` t\n" + // Tables only (1), not views (2), or external tables (3).
    " where type = 1", datasetId.getProject(), datasetId.getDataset(), datasetId.getProject(), datasetId.getDataset());
    TableResult tableRows = bqClient.query(QueryJobConfiguration.newBuilder(tableSql).build());
    List<Callable<BigQueryTable>> tableQueries = new ArrayList<>();
    tableRows.iterateAll().forEach(row -> tableQueries.add(() -> {
        BigQueryTable.Builder table = BigQueryTable.builder().setProject(datasetId.getProject()).setDataset(datasetId.getDataset()).setTableName(row.get(0).getStringValue()).setLastModificationTime(row.get(1).getTimestampValue()).setPartitioningColumn(!row.get(2).isNull() ? row.get(2).getStringValue() : null);
        try {
            if (!loadTableMetadata(table, filter)) {
                return null;
            }
        } catch (RuntimeException e) {
            throw new RuntimeException("Error loading table " + table.getTableName() + " metadata.", e);
        }
        return table.build();
    }));
    ExecutorService executor = Executors.newFixedThreadPool(maxParallelRequests);
    List<Future<BigQueryTable>> tableFutures = executor.invokeAll(tableQueries);
    executor.shutdown();
    List<BigQueryTable> tables = new ArrayList<>(tableFutures.size());
    for (Future<BigQueryTable> ft : tableFutures) {
        BigQueryTable t = ft.get();
        if (t != null) {
            tables.add(t);
        }
    }
    return tables;
}