The user's frustration is understandable, but the error is instructive. MAP and LAMBDA are powerful tools, but they expect a single value per cell in return. TEXTSPLIT, by its nature, returns an array that spills across multiple columns. When MAP feeds each row to TEXTSPLIT, the function tries to push a multi-cell result back into a single cell location, causing the #CALC! error. The logic is sound; the architecture is wrong.
This mistake is a common one when moving from static formulas to dynamic arrays, and it highlights a deeper truth about modern spreadsheet design. The user found an elegant solution with TEXTSPLIT to parse a pattern, numbers.numbers.numbers.text.00000.0000.000000, into clean, separate columns. That is the right instinct. The error is not in the goal, but in the attempt to force a row-by-row approach onto a function built for array output.
The practical fix is to stop thinking in terms of MAP and start thinking in terms of a single array formula. Instead of wrapping TEXTSPLIT in LAMBDA, apply it directly to the entire range. A formula like `=TEXTSPLIT(TEXTJOIN("|", TRUE, B2:B100), ".", "|")` would work, but the cleaner, more modern approach is to use `=DROP(TEXTSPLIT(TEXTJOIN("|", TRUE, B2:B100), ".", "|"), 1)` to skip the header. This allows the function to handle the entire column as one data set, spilling the parsed results across rows and columns automatically.
What this means for you is a shift in mindset. The future of data work in spreadsheets is not about dragging formulas down a column. It is about writing one formula that describes the entire transformation. The pattern the user identified is exactly the kind of structured data that AI-native tools excel at handling. Stop wrestling with individual cells. Let the function see the whole pattern at once. That is where the real productivity gain lives.