读取表格数量差🐠meeeyo.com

Sometimes the number you need isn't in the spreadsheet - it's about the spreadsheet. How many items are in this row? Are two rows the same length? ReadExcelRowOrColumnDiff answers exactly that: it counts the filled cells in a row (or column), optionally compares two of them, and returns the difference as a plain INT.

The Chinese display name is 读取表格数量差, "read the table quantity difference," which is the honest description. It's part of ComfyUI_StringOps, the utility pack that treats Excel as a first-class data source, and it's the arithmetic companion to ReadExcelData - that one gives you the contents, this one gives you the lengths.

How it works

Four inputs:

• excel_path - absolute path to the .xlsx.
• sheet_name - the tab, default Sheet1.
• read_mode - 读行 (read row) or 读列 (read column). These are the actual UI labels; in practice: rows or columns.
• indices - one or two numbers, comma-separated. 1,3 compares index 1 with index 3; a single 2 just counts index 2.

For each index it walks the row or column counting filled cells - and here's the subtle part - it stops at the first empty cell. It's not counting non-empty cells across the whole row; it's measuring a contiguous block. If row 1 is A, B, C, (empty), D, you get a count of 3, not 4. For clean tables (the kind you'd batch-run), that's fine and even desirable; for ragged data it will quietly undercount.

Output is a single INT: the count if you gave one index, or count1 - count2 if you gave two. Use it to drive a comparison node or a conditional - "these two batches don't match, stop."

Where it fits

The realistic use is validation in a batch pipeline. Say you generate one image per product row and one per reference row; wire this node in front of a "if not equal, halt" check so you notice drift before you burn an hour on mismatched batches. It also works as a crude "how long is this list" sensor feeding an integer-based loop or a batch-size calculation.

Like the rest of the pack, errors come back as strings (Error: ...), but here's the trap: this node's output type is INT. If something fails, you get Error: File does not exist at path: ... as a string on an INT port, which will break anything that consumes it. So validate the path up front - the node itself checks existence and read permission and complains helpfully, but you'll see that complaint as an error string, not a number.

Installing it

Standard for the pack: ComfyUI Manager → ComfyUI_StringOps, or

cd ComfyUI/custom_nodes
git clone https://github.com/MeeeyoAI/ComfyUI_StringOps.git

Restart. openpyxl comes from requirements.txt; no models to fetch.

The honest take

This is a narrow node for a specific job - counting contiguous filled cells and diffing two of them. If your sheets are tidy, it's a one-node answer to "did my data get truncated?" If your sheets are messy, the stop-at-first-empty behavior will surprise you, and you might be better off reading the data with ReadExcelData and counting with a string node instead. It's a useful sentry, not a general-purpose Excel tool.