Import `beancount` metadata

[vikasrawal]

Is there support for reading metadata? I can't see the metadata in the beancount file I am reading with ledger. Unless I am missing something.

[trevorld]

• Currently there is no support for reading metadata. Currently we simply generate a csv file using a tuned bean-query command and import that csv data as a data frame.
• What metadata are you interested in?
• Does beancount have a command-line program that outputs that metadata in an easy machine-readable format?

[vikasrawal]

The beancount query can be modified to read the metadata. For example, the following query, reads "document" metadata. Since each user is expected to have have custom metadata tags, it would be desirable to get the user to specify which metadata tags the user wants to import.

paste("select date, flag as mark, account, payee,",
"narration as description,",
"number as amount, currency as commodity,",
"tags,","id,","any_meta('document')")->query

[trevorld]

• Looks like the any_meta("requested_metadata") functionality was added a while ago but isn't documented in the Beancount Query Language document.
• Some examples in this commit: beancount/beancount@30c2b4d .
• If we were to support this the user may need to specify the tags to grab and possibly what column name to save them to (and check they don't conflict with any existing columns).
• We'd also need to update the unit tests to add and test this metadata feature.
• What happens if the key argument for any_meta() is left blank?