Examples of org.kitesdk.data.spi.DatasetRepository

org.kitesdk.data.spi.DatasetRepository

A logical repository (storage system) of {@link Dataset}s.

Implementations of {@code DatasetRepository} are storage systems that containzero or more {@link Dataset}s. A repository acts as a factory, as well as a registry, of datasets. You can {@link #create(String,String,DatasetDescriptor)} anew {@link Dataset} with a name and schema, or retrieve a handle to anexisting dataset, by name, by way of the {@link #load(String,String)} method. Whilenot expressly forbidden, most repositories are expected to support only a single concrete {@link Dataset} implementation.

No guarantees are made as to the durability, reliability, or availability of the underlying storage. That is, a {@code DatasetRepository} could be ondisk, in memory, or some combination. See the implementation class for details about the guarantees it provides.

Implementations of {@link DatasetRepository} are immutable.
@see Dataset @see DatasetDescriptor

    repo.delete("ns", "test");
  }


  @Test
  public void testExternalHDFSQueryOptions() {
    DatasetRepository repo = DatasetRepositories
        .repositoryFor("repo:hive:/tmp/data?" + hdfsQueryArgs);
    repo.delete("ns", "test");
    repo.create("ns", "test", descriptor);


    Dataset<Object> ds = Datasets
        .<Object, Dataset<Object>>load("dataset:hive:/tmp/data/ns/test?" + hdfsQueryArgsOld, Object.class);


    Assert.assertNotNull("Should load dataset", ds);
    Assert.assertTrue(ds instanceof FileSystemDataset);
    Assert.assertEquals("Locations should match",
        URI.create("hdfs://" + hdfsAuth + "/tmp/data/ns/test"),
        ds.getDescriptor().getLocation());
    Assert.assertEquals("Descriptors should match",
        repo.load("ns", "test").getDescriptor(), ds.getDescriptor());


    repo.delete("ns", "test");
  }

View Full Code Here

    repo.delete("ns", "test");
  }


  @Test
  public void testExternalRoot() {
    DatasetRepository repo = DatasetRepositories
        .repositoryFor("repo:hive:/?" + hdfsQueryArgs);
    repo.delete("ns", "test");
    repo.create("ns", "test", descriptor);


    Dataset<Object> ds = Datasets
        .<Object, Dataset<Object>>load("dataset:hive:/ns/test?" + hdfsQueryArgs, Object.class);


    Assert.assertNotNull("Should load dataset", ds);
    Assert.assertTrue(ds instanceof FileSystemDataset);
    Assert.assertEquals("Locations should match",
        URI.create("hdfs://" + hdfsAuth + "/ns/test"),
        ds.getDescriptor().getLocation());
    Assert.assertEquals("Descriptors should match",
        repo.load("ns", "test").getDescriptor(), ds.getDescriptor());
    Assert.assertEquals("Should report correct namespace",
        "ns", ds.getNamespace());
    Assert.assertEquals("Should report correct name",
        "test", ds.getName());


    repo.delete("ns", "test");
  }

View Full Code Here

    repo.delete("ns", "test");
  }


  @Test
  public void testExternalRelative() {
    DatasetRepository repo = DatasetRepositories
        .repositoryFor("repo:hive:data?" + hdfsQueryArgs);
    repo.delete("ns", "test");
    repo.create("ns", "test", descriptor);


    Dataset<Object> ds = Datasets
        .<Object, Dataset<Object>>load("dataset:hive:data/ns/test?" + hdfsQueryArgs, Object.class);


    Assert.assertNotNull("Should load dataset", ds);
    Assert.assertTrue(ds instanceof FileSystemDataset);
    Path cwd = getDFS().makeQualified(new Path("."));
    Assert.assertEquals("Locations should match",
        new Path(cwd, "data/ns/test").toUri(), ds.getDescriptor().getLocation());
    Assert.assertEquals("Descriptors should match",
        repo.load("ns", "test").getDescriptor(), ds.getDescriptor());


    repo.delete("ns", "test");
  }

View Full Code Here

    repo.delete("ns", "test");
  }


  @Test
  public void testManaged() {
    DatasetRepository repo = DatasetRepositories
        .repositoryFor("repo:hive?" + hdfsQueryArgs);
    repo.delete("ns", "test");
    repo.create("ns", "test", descriptor);


    Dataset<Object> ds = Datasets
        .<Object, Dataset<Object>>load("dataset:hive?dataset=test&namespace=ns&" + hdfsQueryArgs, Object.class);


    Assert.assertNotNull("Should load dataset", ds);
    Assert.assertTrue(ds instanceof FileSystemDataset);
    Assert.assertEquals("Descriptors should match",
        repo.load("ns", "test").getDescriptor(), ds.getDescriptor());
    Assert.assertEquals("Should report correct namespace",
        "ns", ds.getNamespace());
    Assert.assertEquals("Should report correct name",
        "test", ds.getName());


    repo.delete("ns", "test");
  }

View Full Code Here

    repo.delete("ns", "test");
  }


  @Test
  public void testManagedDefaultDatabase() {
    DatasetRepository repo = DatasetRepositories
        .repositoryFor("repo:hive?" + hdfsQueryArgs);
    repo.delete("default", "test");
    repo.create("default", "test", descriptor);


    // namespace is not included in the URI
    Dataset<Object> ds = Datasets
        .<Object, Dataset<Object>>load("dataset:hive?dataset=test&" + hdfsQueryArgs, Object.class);


    Assert.assertNotNull("Should load dataset", ds);
    Assert.assertTrue(ds instanceof FileSystemDataset);
    Assert.assertEquals("Descriptors should match",
        repo.load("default", "test").getDescriptor(), ds.getDescriptor());
    Assert.assertEquals("Should report correct namespace",
        "default", ds.getNamespace());
    Assert.assertEquals("Should report correct name",
        "test", ds.getName());


    repo.delete("default", "test");
  }

View Full Code Here

    repo.delete("default", "test");
  }


  @Test
  public void testManagedHDFSQueryOptions() {
    DatasetRepository repo = DatasetRepositories
        .repositoryFor("repo:hive?" + hdfsQueryArgs);
    repo.delete("ns", "test");
    repo.create("ns", "test", descriptor);


    Dataset<Object> ds = Datasets
        .<Object, Dataset<Object>>load("dataset:hive?dataset=test&namespace=ns&" + hdfsQueryArgsOld, Object.class);


    Assert.assertNotNull("Should load dataset", ds);
    Assert.assertTrue(ds instanceof FileSystemDataset);
    Assert.assertEquals("Descriptors should match",
        repo.load("ns", "test").getDescriptor(), ds.getDescriptor());
    Assert.assertEquals("Should report correct namespace",
        "ns", ds.getNamespace());
    Assert.assertEquals("Should report correct name",
        "test", ds.getName());


    repo.delete("ns", "test");
  }

View Full Code Here

    command.datasets = Lists.newArrayList(source, dest);


    int rc = command.run();
    Assert.assertEquals("Should return success", 0, rc);


    DatasetRepository repo = DatasetRepositories.repositoryFor("repo:" + repoUri);
    int size = DatasetTestUtilities.datasetSize(repo.load("default", dest));
    Assert.assertEquals("Should contain copied records", 6, size);


    verify(console).info("Added {} records to \"{}\"", 6l, dest);
    verifyNoMoreInteractions(console);
  }

View Full Code Here

    command.datasets = Lists.newArrayList(source, dest);


    int rc = command.run();
    Assert.assertEquals("Should return success", 0, rc);


    DatasetRepository repo = DatasetRepositories.repositoryFor("repo:" + repoUri);
    Set<GenericRecord> records = DatasetTestUtilities.materialize(
        repo.<GenericRecord>load("default", dest));
    Assert.assertEquals("Should contain copied records", 6, records.size());
    for (GenericRecord record : records) {
      Assert.assertTrue("Username should be upper case",
          UPPER_CASE.matcher(record.get("username").toString()).matches());
    }

View Full Code Here

  @Test
  public void testLoadChangedRelativePathURI() {
    // this used to be a relative external URI, but is now a managed URI
    String uri = "dataset:hive:ns/ds";


    DatasetRepository repo = DatasetRepositories
        .repositoryFor("repo:hive:/tmp/data");
    Dataset<GenericRecord> expected = repo.create(
        "ns", "ds", DESCRIPTOR, GenericRecord.class);


    Dataset<GenericRecord> actual = Datasets.load(uri);
    Assert.assertEquals("Should load existing dataset ns.ds",
        expected, actual);

View Full Code Here

  @Test
  public void testLoadChangedRelativePathURICompatibility() {
    // this used to be a relative external URI, but is now a managed URI
    String uri = "dataset:hive:data/ds";


    DatasetRepository repo = DatasetRepositories
        .repositoryFor("repo:hive:/tmp/data");
    DatasetDescriptor withLocation = new DatasetDescriptor.Builder(DESCRIPTOR)
        .location("file:/tmp/data/ds") // old location
        .build();
    Dataset<GenericRecord> expected = repo.create(
        "default", "ds", withLocation, GenericRecord.class);


    Dataset<GenericRecord> actual = Datasets.load(uri);
    Assert.assertEquals("Should load existing dataset default.ds",
        expected, actual);

View Full Code Here

0 1 2 3 4 5 6 7

TOP

Related Classes of org.kitesdk.data.spi.DatasetRepository

org.kitesdk.cli.commands.DeleteDatasetCommand

org.kitesdk.cli.commands.TestCopyCommandCluster

org.kitesdk.cli.commands.TestCopyCommandLocal

org.kitesdk.cli.commands.TestCreateDatasetCommandCluster

org.kitesdk.cli.commands.TestTransformCommandCluster

org.kitesdk.cli.commands.TestTransformCommandLocal

org.kitesdk.data.Datasets

org.kitesdk.data.hbase.TestHBaseDatasetURIs

org.kitesdk.data.hbase.TestHBaseRepositoryURIs

org.kitesdk.data.mapreduce.DatasetKeyOutputFormat

All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.