why ray.data.read_images cat not combine_chunks

[yanxiaod123]

What happened + What you expected to happen

I use the following code test, i find apply if branch https://github.com/ray-project/ray/blob/master/python/ray/data/_internal/arrow_ops/transform_pyarrow.py#L283，if i have many chunks，how can i to do？because i search #34352 indicate many chunks due to slow.

Versions / Dependencies

ray==3.0.0.dev0

Reproduction script

dataset: Dataset = ray.data.read_images(paths="", size=(224, 224), parallelism=10)
start_time = time.time()
for data in dataset.iter_batches(batch_size=16, batch_format="numpy"):
pass
print('process time is ', time.time() - start_time)

Issue Severity

High: It blocks me from completing my task.