how to find and delete large zotero database entries

psod · March 13, 2017

Firefox, 52.x, is reporting large memory usage to the point where it freezes eventually. The only thing it is doing is Zotero. So I'm assuming that I have added in various documents, like Master's or PhD theses, and I could go without these being loaded all the time. But how to find them, or how to find any document and its file size to be able to think about pruning one's Zotero database? Is it possible to export a list of entries with the size of each?

psod · March 13, 2017

I discovered more information. If I export my library to a CSV format and import that into Excel and sort from small to large on Number of Pages, I can determine the entries with very large sizes (e.g. 19 documents with 6520 pages). However, I also discovered that while Zotero can tell us the number of pages in a PDF from the GUI interface, it doesn't calculate that number on export of the CSV file, so it is necessary to use an spreadsheet formula to determine the total number of pages in one's Journal Article if it has not been reported. Logically, however, the entries reporting Pages are from Journals, and probably are not much more than 30-40 pages in length. The problem Zotero entries will be the records saved in PDF without parent bibliography items. Anyway, it would be useful to output this info somewhere to help others with the same issues. I noticed complaints like mine about this same size issue back to 2010 without action.

adamsmith · March 13, 2017

1) The number of pages is only weakly correlated to the size of a document. You can have 1000 pages on 3MB or two pages with Hi-Resolution images taking up 30+MB.
2) The memory used by Zotero is not at all related to the size of attached files. Zotero doesn't load attachments into memory.
3) If you do need to find large attachments, you can do so pretty quickly with a disk usage analyzer. Sure, it'd occasionally be useful for Zotero to show that information, but that's not easy to do. It's not information available to export scripts, so putting it in the CSV is definitely not an option.
4) I'm not convinced that Zotero is the cause of your Firefox memory issues. It's not impossible, but there are a lot of ways in which memory can leak in Firefox. It's definitely not the case that Zotero, functioning properly, would do this.