ndb

Drop an NDJSON file. Query it.

No schema, no import, no migration.

1. 1

   Point ndb at your files

   Declare a collection in ndb.yml. ndb only reads the files.

   collections:
     events:
       write: direct
       path: /data/logs/*.ndjson
       time: ts

2. 2

   Keep appending lines

   Your app writes one JSON line per event. Any language works.

   echo '{"_id":"4","status":"answered","duration":10,...}' \
     >> /data/logs/app.ndjson

3. 3

   Query and aggregate

   The new line is visible right away, with filters, sort and groups.

   # find where status answered, sort ts -1
   {"_id":"4",...,"status":"answered","duration":10}
   
   # aggregate: $group by status, $sum and $avg
   {"_id":"answered","avg":49,"calls":3}
   {"_id":"missed","avg":0,"calls":1}

• Filters and regex
• Sort and cursor pagination
• Aggregations
• Watch new lines
• Indexes
• Go, Ruby and TypeScript SDKs