JDCN_BatchImageLoadFromList

JDCN_BatchImageLoadFromList is the "page through a list" version of the image loader: instead of scanning a folder itself, it takes a list of paths from somewhere else - JDCN_AnyFileList, a text file, another pack - and loads them in batches, one batch per queue run. If you have 200 images and want to process them 10 at a time across 20 runs, this is the node that turns "a big list" into "the next chunk."

How it works

Four inputs:

• PathList - the paths to load from (list-input).
• Index - which batch to load, 1-based. This is the key mental shift: Index is a page number, not an element offset. With BatchSize = 5, Index = 1 loads items 1–5, Index = 2 loads 6–10. Bump Index, re-queue, get the next page.
• BatchSize - items per batch.
• BatchDirection - TOPTOBOTTOM (front to back), BOTTOMTOTOP (back to front), or RANDOM (random sample of BatchSize from the list).

Outputs: Images (the loaded tensors), ImageNames, ImagePaths (for exactly the items in this batch), and Index echoed back.

How to use it

The intended loop is the standard "chunked processing" pattern:

1. JDCN_AnyFileList scans your folder → PathList.
2. Feed PathList into JDCN_BatchImageLoadFromList with your BatchSize.
3. Queue with Index = 1, then 2, then 3 - each run loads the next slice of the list.
4. ImageNames/ImagePaths let downstream nodes name outputs per batch.

RANDOM direction is a nice touch - random BatchSize-sized slices of your list, handy for variation runs where you don't care about order.

Two honest notes

The loaders expect paths to be real, loadable images. Unlike the path-listing nodes, this one actually opens the files - bad paths get skipped with a console error. And like BatchImageLoadFromDir, each image comes back as its own tensor in a list, so downstream consumers need to handle a list of IMAGE, not a single stacked batch.

Also, Index interacts with BatchSize in the paginated way described above - if you read "Index" as a per-element offset you'll get confused. The pack's own documentation hammers this exact point with the a/b/c/d example: Index = 2, BatchSize = 2 over a 4-element list returns c and d, not b and c.

Installing it

Part of ComfyUI-JDCN:

• ComfyUI Manager → Install Custom Node → search JDCN → install ComfyUI-JDCN → restart.
• Or:

cd ComfyUI/custom_nodes
git clone https://github.com/daxcay/ComfyUI-JDCN.git
cd ComfyUI-JDCN
pip install -r requirements.txt

Restart; under 🔵 JDCN 🔵. Only dependency is piexif, no models.

Common issues

An Index past the end of the list returns an empty batch rather than an error - if you see "no images," check whether you've paged past the last chunk. BatchSize = 0 is accepted by the schema but selects nothing meaningful, so keep it ≥ 1. And if you're pointing it at a PathList that came from a text file, make sure that list is actually a list of paths, one per line - a common copy-paste failure is feeding it raw filenames with no directory prefix.