Convert a MongoDB dump (BSON) to CSV

Read .bson files written by mongodump (or a single BSON document) and turn them into CSV, JSON or any other text, without a running MongoDB.

mongodump --db shop --collection users --out dump
bafi -i dump/shop/users.bson -t users-csv.tmpl -o users.csv

A mongodump file holds one document per record; bafi reads it as a list, so the template iterates with range .. A file with a single document is read as that document.

Try it without MongoDB

Make a small BSON file from JSON with toBSON, then read it back:

{"name": "Alice Smith", "email": "alice@example.com", "roles": ["admin", "billing"]}

bafi -i user.json -t '?{{toBSON .}}' -o user.bson

name,email,roles
"{{.name}}",{{.email}},"{{range $i, $r := .roles}}{{if $i}};{{end}}{{$r}}{{end}}"

bafi -i user.bson -t user-csv.tmpl

name,email,roles
"Alice Smith",alice@example.com,"admin;billing"

For a dump with many users, wrap the data line in {{range .}}...{{end}}.

Inspect a dump

When you do not know the structure of a collection, print the first record:

bafi -i dump/shop/users.bson -t '?{{toJSON (index . 0)}}'

MongoDB's own types (ObjectId, dates) are printed in their Go form; convert them in the template or with toString.

Related: Convert between JSON, YAML, TOON and BSON · JSON to CSV