Local Data

Use this page when teammates need personal test data without merge conflicts.

The two recommended patterns

Local env files

Keep the team baseline committed and personal overrides gitignored:

.env                  # committed — team defaults
.env.secrets          # gitignored — team credentials
.env.local            # gitignored — personal overrides
.env.local.secrets    # gitignored — personal credentials

Switch to personal env files locally with:

glubean run --env-file .env.local

Local JSON data

For structured examples, keep shared and personal files in the same directory:

data/create-user/
  shared.json
  zz-alice.local.json

fromDir() and fromDir.concat() load every matching file and append rows in sorted filename order. fromDir.merge() shallow-merges object files in sorted filename order, so later filenames win on duplicate keys.

There is no special runtime switch for local files. The convention is simpler:

• Keep *.local.json, *.local.yaml, and *.local.yml gitignored.
• Local runs can see personal files that exist on your machine.
• CI sees only committed shared files because gitignored local files are absent.
• If you intentionally want personal values to win for duplicate keys with fromDir.merge(), name the local file so it sorts later, such as zz-alice.local.json.

When you need an explicit shared-only run on your own machine, keep shared and personal data in separate directories or pass an ext/directory choice that matches only the shared files.

Next

• Data-Driven
• CLI Environments & Secrets