Skip to main content
Version: Next (nightly)

Catalogs

Catalogs let you bring in a reference dataset and match your own records against it.

The loop:

  1. Import a dataset as a catalog (a CSV upload, or Pull from source).
  2. Search it from the new-item form, or pick a catalog entry from a record you already have.
  3. The name, image and reference details fill in from the match.

Import a reference set​

Load a dataset (a parts catalog, an ingredient database, a standards list) as a catalog. It sits alongside your own data as something to look things up in, not as records you maintain.

  • CSV upload is how importing works today: name the catalog, upload the file, and its rows become searchable reference entries.
  • Pull from source appears on a catalog that ships with a source URL (the ones a bundle installs usually do). It imports or refreshes its rows in one tap instead of a file upload.
  • Every catalog is browsable: open it to scroll and search its entries directly, like the reference shelf it is.
Very large catalogs shared across an instance

A very large reference catalog (Rebrickable's Lego data, say) does not have to be copied into every workspace. Where the instance holds a shared copy, installing the catalog links to that copy instead, and matching, browsing, and disassembling all read from it. Otherwise it imports into your workspace as usual, and either way the workspace behaves the same.

Match your records to it​

There are two ways in.

  • Match as you create. The new-item form has a search at the top that looks through every installed catalog at once, and picking a result fills in the name, image and reference details before the record even exists. This is the usual way.
  • Match an existing record. Start from a record you already have and pick its catalog entry, and the same details ride along with it.
  • Refresh a URL-built catalog from its source at any time, and Cobb can do it on request. A catalog you built by importing a CSV has no source to refresh from, and it says so rather than pretending to try.

Live API sources and fully automatic matching are planned.