Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I like Excel as much as the next guy, but even in 2013 when you open a delimited text file it makes you manually tell it what the delimiters are. A computer should be able to guess whether a large text file is CSV, TSV, etc. This is an example of something Apple would get right, but Microsoft does not, even after more than a decade with this heavily-used feature. A tasteful implementer would make it do something reasonable by default, rather than making thousands of users click "Next" repeatedly and get presented with nonsensical defaults.


You might be vastly overestimating the correctness of files being imported.

Some coworkers of mine have to deal with CSV/TSV imports from enterprise customers, and the amount of clearly broken scripts is pretty bad.

That excel import preview dialog helps non-programmers try to fix these issues beforehand. If that wasn't there, it'd likely be impossible for these people to know where to begin trying to fix the import.


The preview dialog is useful, but the default option for the delimiter could be guessed and the preview dialog used to confirm it.


I usually just drag tab-sep files into an open Excel window, and it opens just fine. I only get the import wizard when I try to use file -> open. There are a couple of "gotchas" I regularly experience- the worst of which is the automatic formatting- which unfortunately converts a number of gene names into dates (I'm in Bioinformatics). It's enough of an issue that papers have been published about it [1]. For example MAR1 ends up formatted as 1-Mar. When you have a list of 20,000 identifiers, it's quite annoying to deal with. I see the transformed strings show up in biological databases all the time. I don't know why there isn't an easy way to disable autoformatting. Even if I avoid incorporating Excel into my own workflow, when I send documents to scientists, they're going to want to use it, and we end up with the same results then.

[1] http://www.ncbi.nlm.nih.gov/pubmed/15214961


My rage point recently- Why for the love of plain text can Excel not bother to encode csv exports as UTF-8? You can up and then download with google sheets to "hack" around this but I bought numbers.app out of spite since you can specify output encoding there.


When I open a CSV file in Excel 2013, it correctly delimits the file.

You may have non-standard CSV files (or at least delimited files that are not properly escaped)


There are no standard CSV files. There's even no "Excel standard": Excel uses different delimiters in different locales. It's a huge pain.


It's important to point out that there is no standard for CSV files, only conventions.


There is RFC4180, but that came way after people were using it in lots or ways incompatible with the spec, so it isn't a huge help when paraing arbitrary files. I'd recommend following it if you're generating files though.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: