Examples with AssetTable bio.terra.service.dataset.AssetTable used on opensource projects

Example 1 with AssetTable

use of bio.terra.service.dataset.AssetTable in project jade-data-repo by DataBiosphere.

the class SnapshotService method conjureSnapshotTablesFromAsset.

/**
 * Magic up the snapshot tables and snapshot map from the asset tables and columns.
 * This method sets the table lists into snapshot and snapshotSource.
 *
 * @param asset  the one and only asset specification for this snapshot
 * @param snapshot snapshot to point back to and fill in
 * @param snapshotSource snapshotSource to point back to and fill in
 */
private void conjureSnapshotTablesFromAsset(AssetSpecification asset, Snapshot snapshot, SnapshotSource snapshotSource) {
    List<SnapshotTable> tableList = new ArrayList<>();
    List<SnapshotMapTable> mapTableList = new ArrayList<>();
    for (AssetTable assetTable : asset.getAssetTables()) {
        // Create early so we can hook up back pointers.
        SnapshotTable table = new SnapshotTable();
        // Build the column lists in parallel, so we can easily connect the
        // map column to the snapshot column.
        List<Column> columnList = new ArrayList<>();
        List<SnapshotMapColumn> mapColumnList = new ArrayList<>();
        for (AssetColumn assetColumn : assetTable.getColumns()) {
            Column column = new Column(assetColumn.getDatasetColumn());
            columnList.add(column);
            mapColumnList.add(new SnapshotMapColumn().fromColumn(assetColumn.getDatasetColumn()).toColumn(column));
        }
        table.name(assetTable.getTable().getName()).columns(columnList);
        tableList.add(table);
        mapTableList.add(new SnapshotMapTable().fromTable(assetTable.getTable()).toTable(table).snapshotMapColumns(mapColumnList));
    }
    snapshotSource.snapshotMapTables(mapTableList);
    snapshot.snapshotTables(tableList);
}

Example 2 with AssetTable

use of bio.terra.service.dataset.AssetTable in project jade-data-repo by DataBiosphere.

the class BigQueryPdao method queryForRowIds.

// insert the rowIds into the snapshot row ids table and then kick off the rest of the relationship walking
// once we have the row ids in addition to the asset spec, this should look familiar to wAsset
public void queryForRowIds(AssetSpecification assetSpecification, Snapshot snapshot, String sqlQuery) throws InterruptedException {
    BigQueryProject bigQueryProject = bigQueryProjectForSnapshot(snapshot);
    BigQuery bigQuery = bigQueryProject.getBigQuery();
    String snapshotName = snapshot.getName();
    Dataset dataset = snapshot.getSnapshotSources().get(0).getDataset();
    String datasetBqDatasetName = prefixName(dataset.getName());
    String projectId = bigQueryProject.getProjectId();
    // create snapshot bq dataset
    try {
        // create snapshot BQ dataset
        snapshotCreateBQDataset(bigQueryProject, snapshot);
        // now create a temp table with all the selected row ids based on the query in it
        bigQueryProject.createTable(snapshotName, PDAO_TEMP_TABLE, tempTableSchema());
        QueryJobConfiguration queryConfig = QueryJobConfiguration.newBuilder(sqlQuery).setDestinationTable(TableId.of(snapshotName, PDAO_TEMP_TABLE)).setWriteDisposition(JobInfo.WriteDisposition.WRITE_APPEND).build();
        try {
            final TableResult query = bigQuery.query(queryConfig);
            // get results and validate that it got back more than 0 value
            if (query.getTotalRows() < 1) {
                // should this be a different error?
                throw new InvalidQueryException("Query returned 0 results");
            }
        } catch (InterruptedException ie) {
            throw new PdaoException("Append query unexpectedly interrupted", ie);
        }
        // join on the root table to validate that the dataset's rootTable.rowid is never null
        // and thus matches the PDAO_ROW_ID_COLUMN
        AssetTable rootAssetTable = assetSpecification.getRootTable();
        Table rootTable = rootAssetTable.getTable();
        String datasetTableName = rootTable.getName();
        String rootTableId = rootTable.getId().toString();
        ST sqlTemplate = new ST(joinTablesToTestForMissingRowIds);
        sqlTemplate.add("snapshotDatasetName", snapshotName);
        sqlTemplate.add("tempTable", PDAO_TEMP_TABLE);
        sqlTemplate.add("datasetDatasetName", datasetBqDatasetName);
        sqlTemplate.add("datasetTable", datasetTableName);
        sqlTemplate.add("commonColumn", PDAO_ROW_ID_COLUMN);
        TableResult result = bigQueryProject.query(sqlTemplate.render());
        FieldValueList mismatchedCount = result.getValues().iterator().next();
        Long mismatchedCountLong = mismatchedCount.get(0).getLongValue();
        if (mismatchedCountLong > 0) {
            throw new MismatchedValueException("Query results did not match dataset root row ids");
        }
        // TODO should this be pulled up to the top of queryForRowIds() / added to snapshotCreateBQDataset() helper
        bigQueryProject.createTable(snapshotName, PDAO_ROW_ID_TABLE, rowIdTableSchema());
        // populate root row ids. Must happen before the relationship walk.
        // NOTE: when we have multiple sources, we can put this into a loop
        // insert into the PDAO_ROW_ID_TABLE the literal that is the table id
        // and then all the row ids from the temp table
        ST sqlLoadTemplate = new ST(loadRootRowIdsFromTempTableTemplate);
        sqlLoadTemplate.add("project", projectId);
        sqlLoadTemplate.add("snapshot", snapshotName);
        sqlLoadTemplate.add("dataset", datasetBqDatasetName);
        sqlLoadTemplate.add("tableId", rootTableId);
        // this is the disc from classic asset
        sqlLoadTemplate.add("commonColumn", PDAO_ROW_ID_COLUMN);
        sqlLoadTemplate.add("tempTable", PDAO_TEMP_TABLE);
        bigQueryProject.query(sqlLoadTemplate.render());
        // ST sqlValidateTemplate = new ST(validateRowIdsForRootTemplate);
        // TODO do we want to reuse this validation? if yes, maybe mismatchedCount / query should be updated
        // walk and populate relationship table row ids
        List<WalkRelationship> walkRelationships = WalkRelationship.ofAssetSpecification(assetSpecification);
        walkRelationships(datasetBqDatasetName, snapshotName, walkRelationships, rootTableId, projectId, bigQuery);
        // populate root row ids. Must happen before the relationship walk.
        // NOTE: when we have multiple sources, we can put this into a loop
        snapshotViewCreation(datasetBqDatasetName, snapshotName, snapshot, projectId, bigQuery, bigQueryProject);
    } catch (PdaoException ex) {
        // TODO what if the query is invalid? Seems like there might be more to catch here.
        throw new PdaoException("createSnapshot failed", ex);
    }
}