Questions about this extension

[chitralverma]

Hi thanks for the extension, had some questions about it.

• Does this support lazy IO?
• Does this support streaming read or is all the data loaded in mem?
• Does this rely on httpfs extension to read from remote stores (s3, gcs, ... )?
• Does this support glob paths?
• How does this compare against the other extension?
• Is this still being maintained?

ref: duckdb/duckdb-excel#48

[freddie-freeloader]

Hey @chitralverma ☺

• Does this support lazy IO?
• Does this support streaming read or is all the data loaded in mem?

This extension builds on the parser described in this paper: Information Systems Journal
The file is parsed as a whole, but it will not build a full XML-tree -- so it's rather memory efficient.

• Does this rely on httpfs extension to read from remote stores (s3, gcs, ... )?

This extension accepts a file path to the file system -- so I think this wouldn't work.

• Does this support glob paths?

Only one sheet of one file can be parsed at a time & we don't provide support for defining multiple sheets/files as input.

• How does this compare against the other extension?

This extension was build, when only the spatial extension was able to read .xlsx-files -- this is why our README contains a benchmark against this one. This new extension, (released with DuckDB v1.2) does seem to be a replacement -- I haven't used it so far -- so I cannot tell you in detail how the two compare.

• Is this still being maintained?

This was created as part of a uni project, so it's not actively developed -- if you already use it & rely on it + found any issues, I would definitely try to fix them. If you don't rely on this extension already & there are not any features, that this extension uniquely provides, the official extension is probably the more reliable choice. 😊

I hope this answers your questions. ☺
Kind regards from Berlin