min

Need to know the smallest value sitting in a list of numbers? That's the whole job of DataListMin. It takes a list of FLOAT or INT values, finds the lowest one, and hands it back. If you've seen DataListMax, this is its mirror image - same shape, opposite goal.

One input (list), one output (min), and the returned value keeps the type of whatever it found. A list of ints gives you an int, floats give you a float, and a mixed list gives you whichever type the winning value happened to have.

How it works

Under the hood it's Python's built-in min() over the whole collection, with two guard rails that make it behave nicely inside a graph. An empty list returns None instead of raising. A list containing non-numeric or non-comparable values also returns None instead of crashing the workflow with a TypeError. In ComfyUI, a node that dies mid-run takes the whole graph down with it, so "return nothing and let the next node cope" is the right call for a utility node.

Where it earns its keep

The classic move is range checking: is anything in this batch below the threshold? Wire min into the pack's comparison nodes and you can branch on "everything's above the floor" in one step. It's also half of any manual normalization - min and max together give you the bounds of your data, and then a little division rescales everything to a 0–1 range. The pack doesn't give you a dedicated normalize node, but min + max + a math node gets you there without leaving the pack.

The rest of the family is worth knowing: DataListMax is the ceiling, DataListSum totals the list, and DataListSort shows you the whole ordering if you need more than the single extreme value.

Installing it

This is one of 250+ nodes in StableLlama's Basic data handling pack, which is pure Python with zero third-party dependencies and no model files. Install the pack once and this node (and everything else) is yours:

cd ComfyUI/custom_nodes
git clone https://github.com/StableLlama/ComfyUI-basic_data_handling

Restart ComfyUI after cloning. Or, the easy route: search "Basic data handling" in ComfyUI Manager and click install. Because the pack has no requirements.txt of its own, you sidestep the dependency-conflict roulette that plagues heavier custom-node packs.

Common issues

Same family of gotchas as max: the input is typed FLOAT,INT, so strings slide through to the None fallback rather than being converted - run them through a cast node first. And don't confuse the empty-list None with a real number. If you wired min into a math node and it's quietly producing nothing, check whether the upstream list actually has items in it before you suspect the node.