use of bio.terra.grammar.exception.InvalidQueryException in project jade-data-repo by DataBiosphere.
the class Query method parse.
public static Query parse(String sql) {
CharStream charStream = CharStreams.fromString(sql);
SQLLexer lexer = new SQLLexer(charStream);
SQLParser parser = new SQLParser(new CommonTokenStream(lexer));
parser.setErrorHandler(new BailErrorStrategy());
try {
return new Query(parser, parser.query_statement());
} catch (ParseCancellationException ex) {
throw new InvalidQueryException("Could not parse query: " + sql, ex);
}
}
use of bio.terra.grammar.exception.InvalidQueryException 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);
}
}
use of bio.terra.grammar.exception.InvalidQueryException in project jade-data-repo by DataBiosphere.
the class CreateSnapshotPrimaryDataQueryStep method doStep.
@Override
public StepResult doStep(FlightContext context) throws InterruptedException {
// TODO: this assumes single-dataset snapshots, will need to add a loop for multiple
// (based on the validation flight step that already occurred.)
/*
* get dataset and assetName
* get asset from dataset
* which gives the root table
* to use in conjunction with the filtered row ids to create this snapshot
*/
Snapshot snapshot = snapshotDao.retrieveSnapshotByName(snapshotReq.getName());
SnapshotRequestQueryModel snapshotQuerySpec = snapshotReq.getContents().get(0).getQuerySpec();
String snapshotAssetName = snapshotQuerySpec.getAssetName();
String snapshotQuery = snapshotReq.getContents().get(0).getQuerySpec().getQuery();
Query query = Query.parse(snapshotQuery);
List<String> datasetNames = query.getDatasetNames();
// TODO this makes the assumption that there is only one dataset
// (based on the validation flight step that already occurred.)
// This will change when more than 1 dataset is allowed
String datasetName = datasetNames.get(0);
Dataset dataset = datasetService.retrieveByName(datasetName);
DatasetModel datasetModel = datasetService.retrieveModel(dataset);
// get asset out of dataset
Optional<AssetSpecification> assetSpecOp = dataset.getAssetSpecificationByName(snapshotAssetName);
AssetSpecification assetSpec = assetSpecOp.orElseThrow(() -> new AssetNotFoundException("Expected asset specification"));
Map<String, DatasetModel> datasetMap = Collections.singletonMap(datasetName, datasetModel);
BigQueryVisitor bqVisitor = new BigQueryVisitor(datasetMap);
String sqlQuery = query.translateSql(bqVisitor);
// validate that the root table is actually a table being queried in the query -->
// and the grammar only picks up tables names in the from clause (though there may be more than one)
List<String> tableNames = query.getTableNames();
String rootTablename = assetSpec.getRootTable().getTable().getName();
if (!tableNames.contains(rootTablename)) {
throw new InvalidQueryException("The root table of the selected asset is not present in this query");
}
// now using the query, get the rowIds
// insert the rowIds into the snapshot row ids table and then kick off the rest of the relationship walking
bigQueryPdao.queryForRowIds(assetSpec, snapshot, sqlQuery);
return StepResult.getStepResultSuccess();
}
Aggregations