use of co.cask.cdap.api.dataset.table.Get in project cdap by caskdata.
the class IndexedTableTest method testIndexedOperations.
@Test
public void testIndexedOperations() throws Exception {
TransactionExecutor txnl = dsFrameworkUtil.newTransactionExecutor(table);
// start a new transaction
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
// add a value c with idx = 1, and b with idx = 2
table.put(new Put(keyC).add(idxCol, idx1).add(valCol, valC));
table.put(new Put(keyB).add(idxCol, idx2).add(valCol, valB));
}
});
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
// read by key c
Row row = table.get(new Get(keyC, colIdxVal));
TableAssert.assertColumns(row, colIdxVal, new byte[][] { idx1, valC });
// read by key b
row = table.get(new Get(keyB, colIdxVal));
TableAssert.assertColumns(row, colIdxVal, new byte[][] { idx2, valB });
// read by idx 1 -> c
row = readFirst(table.readByIndex(idxCol, idx1));
TableAssert.assertColumns(row, colIdxVal, new byte[][] { idx1, valC });
// read by idx 2 -> b
row = readFirst(table.readByIndex(idxCol, idx2));
TableAssert.assertColumns(row, colIdxVal, new byte[][] { idx2, valB });
// test read over empty index (idx 3)
assertEmpty(table.readByIndex(idxCol, idx3));
}
});
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
// add a value a with idx = 1
table.put(new Put(keyA).add(idxCol, idx1).add(valCol, valA));
}
});
// read by idx 1 -> a
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
Row row = readFirst(table.readByIndex(idxCol, idx1));
TableAssert.assertColumns(row, colIdxVal, new byte[][] { idx1, valA });
}
});
// start a new transaction
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
// delete value a
table.delete(new Delete(keyA, colIdxVal));
}
});
// read by idx 1 -> c
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
Row row = readFirst(table.readByIndex(idxCol, idx1));
TableAssert.assertColumns(row, colIdxVal, new byte[][] { idx1, valC });
}
});
// start a new transaction
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
// add a value aa with idx 2
table.put(new Put(keyAA).add(idxCol, idx2).add(valCol, valAA));
}
});
// read by idx 2 -> aa
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
Row row = readFirst(table.readByIndex(idxCol, idx2));
TableAssert.assertColumns(row, colIdxVal, new byte[][] { idx2, valAA });
}
});
// start a new transaction
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
// swap value for aa to ab
Assert.assertTrue(table.compareAndSwap(keyAA, valCol, valAA, valAB));
}
});
// read by idx 2 -> ab
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
Row row = readFirst(table.readByIndex(idxCol, idx2));
TableAssert.assertColumns(row, colIdxVal, new byte[][] { idx2, valAB });
}
});
// start a new transaction
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
// swap value for aa to bb
Assert.assertTrue(table.compareAndSwap(keyAA, valCol, valAB, valBB));
}
});
// read by idx 2 -> bb (value of key aa)
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
Row row = readFirst(table.readByIndex(idxCol, idx2));
TableAssert.assertColumns(row, colIdxVal, new byte[][] { idx2, valBB });
}
});
// start a new transaction
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
// swap value for aa to null
Assert.assertTrue(table.compareAndSwap(keyAA, valCol, valBB, null));
}
});
// read by idx 2 -> null (value of b)
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
Row row = readFirst(table.readByIndex(idxCol, idx2));
TableAssert.assertColumn(row, idxCol, idx2);
}
});
// start a new transaction
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
// swap idx for c to 3
Assert.assertTrue(table.compareAndSwap(keyC, idxCol, idx1, idx3));
}
});
// read by idx 1 -> null (no row has that any more)
txnl.execute(new TransactionExecutor.Subroutine() {
@Override
public void apply() throws Exception {
assertEmpty(table.readByIndex(idxCol, idx1));
// read by idx 3 > c
Row row = readFirst(table.readByIndex(idxCol, idx3));
TableAssert.assertColumns(row, new byte[][] { idxCol, valCol }, new byte[][] { idx3, valC });
}
});
}
use of co.cask.cdap.api.dataset.table.Get in project cdap by caskdata.
the class ProgramScheduleStoreDataset method findSchedules.
/**
* Find all schedules that have a trigger with a given trigger key.
*
* @param triggerKey the trigger key to look up
* @return a list of all schedules that are triggered by this key; never null
*/
public Collection<ProgramScheduleRecord> findSchedules(String triggerKey) {
Map<ScheduleId, ProgramScheduleRecord> schedulesFound = new HashMap<>();
try (Scanner scanner = store.readByIndex(TRIGGER_KEY_COLUMN_BYTES, Bytes.toBytes(triggerKey))) {
Row triggerRow;
while ((triggerRow = scanner.next()) != null) {
String triggerRowKey = Bytes.toString(triggerRow.getRow());
try {
ScheduleId scheduleId = extractScheduleIdFromTriggerKey(triggerRowKey);
if (schedulesFound.containsKey(scheduleId)) {
continue;
}
Row row = store.get(new Get(rowKeyForSchedule(scheduleId)));
byte[] serialized = row.get(SCHEDULE_COLUMN_BYTES);
if (serialized == null) {
throw new NotFoundException(scheduleId);
}
ProgramSchedule schedule = GSON.fromJson(Bytes.toString(serialized), ProgramSchedule.class);
ProgramScheduleMeta meta = extractMetaFromRow(scheduleId, row);
ProgramScheduleRecord record = new ProgramScheduleRecord(schedule, meta);
schedulesFound.put(scheduleId, record);
} catch (IllegalArgumentException | NotFoundException e) {
// the only exceptions we know to be thrown here are IllegalArgumentException (ill-formed key) or
// NotFoundException (if the schedule does not exist). Both should never happen, so we warn and ignore.
// we will let any other exception propagate up, because it would be a DataSetException or similarly serious.
LOG.warn("Problem with trigger id '{}' found for trigger key '{}': {}. Skipping entry.", triggerRowKey, triggerKey, e.getMessage());
}
}
}
return schedulesFound.values();
}
use of co.cask.cdap.api.dataset.table.Get in project cdap by caskdata.
the class ProgramScheduleStoreDataset method getSchedule.
/**
* Read a schedule from the store.
*
* @param scheduleId the id of the schedule to read
* @return the schedule from the store
* @throws NotFoundException if the schedule does not exist in the store
*/
public ProgramSchedule getSchedule(ScheduleId scheduleId) throws NotFoundException {
Row row = store.get(new Get(rowKeyForSchedule(scheduleId)));
byte[] serialized = row.get(SCHEDULE_COLUMN_BYTES);
if (serialized == null) {
throw new NotFoundException(scheduleId);
}
return GSON.fromJson(Bytes.toString(serialized), ProgramSchedule.class);
}
use of co.cask.cdap.api.dataset.table.Get in project cdap by caskdata.
the class BufferingTableTest method testMultiGetIncludesBuffer.
@Test
public void testMultiGetIncludesBuffer() throws Exception {
DatasetAdmin admin = getTableAdmin(CONTEXT1, MY_TABLE);
admin.create();
try {
// persist some data
BufferingTable table = getTable(CONTEXT1, MY_TABLE);
Transaction tx1 = txClient.startShort();
table.startTx(tx1);
// writing a couple rows
// table should look like the following, with everything in the buffer
// c1 c2 c3 c4
// r1 1 2 3 -
// r2 - 3 2 1
table.put(R1, a(C1, C2, C3), lb(1, 2, 3));
table.put(R2, a(C2, C3, C4), lb(3, 2, 1));
// check that multi-get can see buffered writes
List<Row> rows = table.get(Lists.newArrayList(new Get(R1), new Get(R2)));
Assert.assertEquals(2, rows.size());
TableAssert.assertRow(rows.get(0), R1, a(C1, C2, C3), lb(1, 2, 3));
TableAssert.assertRow(rows.get(1), R2, a(C2, C3, C4), lb(3, 2, 1));
// check multi-get with gets that specify columns, and one get that should return an empty row
rows = table.get(Lists.newArrayList(new Get(R1, C2, C3), new Get(R2, C2, C3), new Get(R3)));
Assert.assertEquals(3, rows.size());
TableAssert.assertRow(rows.get(0), R1, a(C2, C3), lb(2, 3));
TableAssert.assertRow(rows.get(1), R2, a(C2, C3), lb(3, 2));
Assert.assertTrue(rows.get(2).isEmpty());
// persist changes
Collection<byte[]> txChanges = table.getTxChanges();
Assert.assertTrue(txClient.canCommit(tx1, txChanges));
Assert.assertTrue(table.commitTx());
Assert.assertTrue(txClient.commit(tx1));
table.postTxCommit();
// start another transaction
Transaction tx2 = txClient.startShort();
table.startTx(tx2);
// now add another row, delete a row, and change some column values
// table should look like the following
// c1 c2 c3 c4 c5
// r1 - - 3 2 -
// r3 - - - - 1
table.put(R1, a(C2, C3, C4), lb(4, 3, 2));
table.delete(R1, a(C1, C2));
table.delete(R2);
table.put(R3, C5, L1);
// verify multi-get sees persisted data with buffer applied on top
rows = table.get(Lists.newArrayList(new Get(R1), new Get(R2), new Get(R3)));
Assert.assertEquals(3, rows.size());
TableAssert.assertRow(rows.get(0), R1, a(C3, C4), lb(3, 2));
Assert.assertTrue(rows.get(1).isEmpty());
TableAssert.assertRow(rows.get(2), R3, a(C5), lb(1));
// pretend there was a write conflict and rollback changes
Assert.assertTrue(table.rollbackTx());
txClient.abort(tx2);
// start another transaction and make sure it can't see what was done before
Transaction tx3 = txClient.startShort();
table.startTx(tx3);
rows = table.get(Lists.newArrayList(new Get(R1), new Get(R2)));
Assert.assertEquals(2, rows.size());
TableAssert.assertRow(rows.get(0), R1, a(C1, C2, C3), lb(1, 2, 3));
TableAssert.assertRow(rows.get(1), R2, a(C2, C3, C4), lb(3, 2, 1));
} finally {
admin.drop();
}
}
use of co.cask.cdap.api.dataset.table.Get in project cdap by caskdata.
the class TopKCollector method readWordAssocs.
/**
* Returns the top words associated with the specified word and the number
* of times the words have appeared together.
* @param word the word of interest
* @param limit the number of associations to return, at most
* @return a map of the top associated words to their co-occurrence count
*/
@ReadOnly
public Map<String, Long> readWordAssocs(String word, int limit) {
// Retrieve all columns of the word’s row
Row result = this.table.get(new Get(word));
TopKCollector collector = new TopKCollector(limit);
if (!result.isEmpty()) {
// Iterate over all columns
for (Map.Entry<byte[], byte[]> entry : result.getColumns().entrySet()) {
collector.add(Bytes.toLong(entry.getValue()), Bytes.toString(entry.getKey()));
}
}
return collector.getTopK();
}
Aggregations