I want to make an app that allows folks to upload an excel file and I want to extract just the top 10k rows of the first two columns to plot them. I’m afraid of allowing an upload option for a free service, can I parse an xls file using wasm and avoid uploading anything?
For client-side stuff like that, there is sheetjs which has support for XLSX but there are file size limitations. XLSX is a glorified zip file and to parse it you 1st have to unzip it and then build the XLSX file.
I found that it starts to hang on XLSX files ~50MB because the spreadsheet model it produces get's a bit too big and sheetjs dies. Doing this in a web worker is a good idea because otherwise it will clog the main thread.
For CSV the story is different, because you can read a plain text file in chunks (even in JS) with FileReader and readAsArrayBuffer(file.slice(byteOffset, byteOffset + chunkSize, file.type)).
