A 10-yuan Taobao Excel batch-processing gig—I ran it myself with WorkBuddy

Last Wednesday night working overtime, I searched "excel batch processing" on Taobao, and the top 100 items popped up. VBA gigs for 10 yuan, Python scripts for 100 yuan, table merge mini-tools for 9.99, all 5.0 stars. I stared at the screen for two minutes. Right beside me were 27 reimbursement detail sheets submitted by departments, with a summary due the next morning.

excel data processing analysis cleaning organizing entry extraction merge statistics comparison batch gig VBA service ¥ 10

I started running this task with WorkBuddy a month ago, and now I basically don't hire anyone. Below is the full method, for first-timers' reference.

27 sheets, dates in three formats, amounts sometimes with ¥ sometimes plain numbers, department names sometimes with spaces sometimes abbreviated. The problem with hiring someone is they write you a script, and next time the headers change you have to start over.

The WorkBuddy path is: open the desktop app, left sidebar workbench, new task, drag the 27 files directly into the upload area. It scans first, gives you a field list, asks which column is date, which is amount, which is department. This step is called field mapping—telling it what your column is called and which column in my sheet it corresponds to. Don't skip it out of impatience; skip it and everything goes haywire later. I usually set date format to yyyy-mm-dd and strip currency symbols from amounts here.

Merging is just the first step. After merging, I have it dedupe by invoice number, then compare against already-reimbursed records, outputting an anomaly list. I mainly use that latter list.

Here's something I misjudged before. I used to think WorkBuddy was okay at extracting invoice fields, but couldn't integrate with Meike, so the whole chain couldn't be built. Later I figured it out—I don't need it to connect the whole chain, I just need it to pick out suspicious rows, and I verify them myself. This is also safer, data doesn't need to go out.

On config: output format xlsx, check keep original row numbers, add a column for anomaly reason. The key parameter is the duplicate threshold—I set it so only identical invoice numbers count as duplicates, same amount doesn't, because restaurant receipts hitting the same amount is too common.

Here's the pitfall. First run, it flagged two reimbursements of the same invoice as duplicates, but they were actually normal supplementary filings. Later I added a line in the task description: same invoice number within 7 days counts as duplicate. You have to spell out these rules yourself; it won't guess for you.

Permissions: I set the source folder to read-only, output written to a separate shared folder, colleagues can only view not modify. Financial data must be locked down on this point.

There's an article on Taobao Baike about office file batch processing, saying that in 2026 out-of-the-box usability is becoming the new threshold. I quite agree. The trouble with hiring script gigs is you have to rehire whenever the sheet changes; WorkBuddy is drag-your-own-files, edit-your-own-rules, and it runs on the spot after editing.

Daily maintenance is just two things. Run once on the 25th of each month; before running, clear old files from the source folder, otherwise it reads last month's in too. I stepped on this pit once—thirty-some extra rows, only noticed during reconciliation.

My environment is Windows desktop, output to shared drive; haven't tried on Mac, may not apply to everyone.

Now after running each month, I just take that anomaly list to verify; nobody even looks at the merged table itself.