use of org.apache.hyracks.storage.am.common.dataflow.IndexCreateOperatorDescriptor in project asterixdb by apache.
the class DatasetUtil method createDatasetJobSpec.
public static JobSpecification createDatasetJobSpec(Dataset dataset, MetadataProvider metadataProvider) throws AlgebricksException {
Index index = IndexUtil.getPrimaryIndex(dataset);
ARecordType itemType = (ARecordType) metadataProvider.findType(dataset);
// get meta item type
ARecordType metaItemType = null;
if (dataset.hasMetaPart()) {
metaItemType = (ARecordType) metadataProvider.findMetaType(dataset);
}
JobSpecification spec = RuntimeUtils.createJobSpecification(metadataProvider.getApplicationContext());
Pair<IFileSplitProvider, AlgebricksPartitionConstraint> splitsAndConstraint = metadataProvider.getSplitProviderAndConstraints(dataset);
FileSplit[] fs = splitsAndConstraint.first.getFileSplits();
StringBuilder sb = new StringBuilder();
for (int i = 0; i < fs.length; i++) {
sb.append(fs[i] + " ");
}
LOGGER.info("CREATING File Splits: " + sb.toString());
Pair<ILSMMergePolicyFactory, Map<String, String>> compactionInfo = DatasetUtil.getMergePolicyFactory(dataset, metadataProvider.getMetadataTxnContext());
//prepare a LocalResourceMetadata which will be stored in NC's local resource repository
IResourceFactory resourceFactory = dataset.getResourceFactory(metadataProvider, index, itemType, metaItemType, compactionInfo.first, compactionInfo.second);
IndexBuilderFactory indexBuilderFactory = new IndexBuilderFactory(metadataProvider.getStorageComponentProvider().getStorageManager(), splitsAndConstraint.first, resourceFactory, !dataset.isTemp());
IndexCreateOperatorDescriptor indexCreateOp = new IndexCreateOperatorDescriptor(spec, indexBuilderFactory);
AlgebricksPartitionConstraintHelper.setPartitionConstraintInJobSpec(spec, indexCreateOp, splitsAndConstraint.second);
spec.addRoot(indexCreateOp);
return spec;
}
Aggregations