pabloem commented on a change in pull request #15863:
URL: https://github.com/apache/beam/pull/15863#discussion_r748533384
##########
File path:
sdks/java/io/jdbc/src/main/java/org/apache/beam/sdk/io/jdbc/JdbcIO.java
##########
@@ -1494,8 +1509,49 @@ void set(
checkArgument(
(getDataSourceProviderFn() != null),
"withDataSourceConfiguration() or withDataSourceProviderFn() is
required");
-
- return input.apply(
+ checkArgument(
+ ((getAutoSharding() == null || !getAutoSharding())
Review comment:
hehe thanks! appreciate it.
##########
File path:
sdks/java/io/jdbc/src/main/java/org/apache/beam/sdk/io/jdbc/JdbcIO.java
##########
@@ -1494,8 +1509,49 @@ void set(
checkArgument(
(getDataSourceProviderFn() != null),
"withDataSourceConfiguration() or withDataSourceProviderFn() is
required");
-
- return input.apply(
+ checkArgument(
+ ((getAutoSharding() == null || !getAutoSharding())
+ && input.isBounded() == IsBounded.BOUNDED)
+ || input.isBounded() == IsBounded.UNBOUNDED,
+ "Autosharding is only supported for streaming pipelines.");
+ ;
+
+ PCollection<Iterable<T>> iterables;
+ if (input.isBounded() == IsBounded.UNBOUNDED
+ && getAutoSharding() != null
+ && getAutoSharding()) {
+ iterables =
+ input
+ .apply(WithKeys.<String, T>of(""))
+ .apply(
+ GroupIntoBatches.<String, T>ofSize(DEFAULT_BATCH_SIZE)
+ .withMaxBufferingDuration(Duration.millis(200))
+ .withShardedKey())
+ .apply(Values.create());
+ } else {
+ iterables =
+ input.apply(
+ ParDo.of(
+ new DoFn<T, Iterable<T>>() {
+ List<T> outputList;
+
+ @ProcessElement
+ public void process(ProcessContext c) {
+ if (outputList == null) {
+ outputList = new ArrayList<>();
+ }
+ outputList.add(c.element());
+ }
+
+ @FinishBundle
+ public void finish(FinishBundleContext c) {
+ System.out.println("List size is " +
outputList.size());
Review comment:
removed. Thanks!
##########
File path:
sdks/java/io/jdbc/src/main/java/org/apache/beam/sdk/io/jdbc/JdbcIO.java
##########
@@ -1494,8 +1509,49 @@ void set(
checkArgument(
(getDataSourceProviderFn() != null),
"withDataSourceConfiguration() or withDataSourceProviderFn() is
required");
-
- return input.apply(
+ checkArgument(
+ ((getAutoSharding() == null || !getAutoSharding())
+ && input.isBounded() == IsBounded.BOUNDED)
+ || input.isBounded() == IsBounded.UNBOUNDED,
+ "Autosharding is only supported for streaming pipelines.");
+ ;
+
+ PCollection<Iterable<T>> iterables;
+ if (input.isBounded() == IsBounded.UNBOUNDED
+ && getAutoSharding() != null
+ && getAutoSharding()) {
+ iterables =
+ input
+ .apply(WithKeys.<String, T>of(""))
+ .apply(
+ GroupIntoBatches.<String, T>ofSize(DEFAULT_BATCH_SIZE)
+ .withMaxBufferingDuration(Duration.millis(200))
+ .withShardedKey())
+ .apply(Values.create());
+ } else {
+ iterables =
Review comment:
that's correct. Since the WriteFn now takes Lists instead of individual
elements, we need to perform batching in the previous operation
##########
File path:
sdks/java/io/jdbc/src/main/java/org/apache/beam/sdk/io/jdbc/JdbcIO.java
##########
@@ -1432,6 +1442,11 @@ void set(
return toBuilder().setPreparedStatementSetter(setter).build();
}
+ /** If true, enables using a dynamically determined number of shards to
write. */
+ public WriteWithResults<T, V> withAutoSharding() {
Review comment:
hmmm I think we should wait for feedback before adding that capability.
If users find that they want more control over sharding, they'd probably file
issues / ask questions on stackoverflow. If we add this capability we'll need
to support it for the longer term... What do you think?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
