imbajin commented on code in PR #683:
URL:
https://github.com/apache/incubator-hugegraph-toolchain/pull/683#discussion_r2444650259
##########
hugegraph-loader/src/main/java/org/apache/hugegraph/loader/HugeGraphLoader.java:
##########
@@ -200,27 +594,133 @@ private void loadInputs(List<InputStruct> structs) {
}
}
- private void loadStructs(List<InputStruct> structs) {
- // Load input structs one by one
+ private List<InputTaskItem> prepareTaskItems(List<InputStruct> structs,
+ boolean scatter) {
+ ArrayList<InputTaskItem> tasks = new ArrayList<>();
+ int curFile = 0;
+ int curIndex = 0;
for (InputStruct struct : structs) {
- if (this.context.stopped()) {
- break;
- }
if (struct.skip()) {
continue;
}
- // Create and init InputReader, fetch next batch lines
- try (InputReader reader = InputReader.create(struct.input())) {
- // Init reader
- reader.init(this.context, struct);
- // Load data from current input mapping
- this.loadStruct(struct, reader);
+
+ // Create and init InputReader
+ try {
+ LOG.info("Start loading: '{}'", struct);
+
+ InputReader reader = InputReader.create(struct.input());
+ List<InputReader> readerList = reader.multiReaders() ?
+ reader.split() :
+ ImmutableList.of(reader);
+
+ LOG.info("total {} found in '{}'", readerList.size(), struct);
+ tasks.ensureCapacity(tasks.size() + readerList.size());
+ int seq = 0;
+ for (InputReader r : readerList) {
+ if (curFile >= this.context.options().startFile &&
+ (this.context.options().endFile == -1 ||
+ curFile < this.context.options().endFile)) {
+ // Load data from current input mapping
+ tasks.add(new InputTaskItem(struct, r, seq, curIndex));
Review Comment:
**Resource Cleanup Issue**: When readers are filtered out by the file range
check (startFile/endFile), they are closed with `r.close()`. However, if an
exception occurs during the loop before reaching these readers, earlier readers
won't be cleaned up. Consider tracking all created readers in a list and using
try-finally to ensure cleanup.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
