Mongo-Postgres Query Converter

MongoDB query documents are quite powerful. This brings that usefulness to PostgreSQL by letting you query in a similar way. This tool converts a Mongo query to a PostgreSQL where clause for data stored in a jsonb field. It also has additional converters for Mongo projections which are like select clauses and for update queries.

This tool is used by pgmongo which intends to provide a drop-in replacement for MongoDB.

Installation

npm install mongo-query-to-postgres-jsonb

Simple Usage

var mToPsql = require('mongo-query-to-postgres-jsonb')
var query = { field: 'value' }
var sqlQuery = mToPsql('data', query)

API

var mToPsql = require('mongo-query-to-postgres-jsonb')

mToPsql(sqlField, mongoQuery, [arrayFields])

sqlField

This is the name of your jsonb column in your postgres table which holds all the data.

mongoQuery

An object containing MongoDB query operators.

arrayFields

This tool doesn't know which fields are arrays so you can optionally specify a list of dotted paths which should be treated as an array.

mToPsql.convertSelect(sqlField, projectionQuery, [arrayFields])

projectionQuery

Object specifying which a subset of documents to return. Note: advanced projection fields are not yet supported.

mToPsql.convertUpdate(sqlField, updateQuery, [upsert])

updateQuery

Object containing MongoDB operations to apply to the documents.

upsert

Indicate that the query is being used for upserting. This will create a safer query that works if the original document doesn't already exist.

mToPsql.convertSort(sqlField, sortQuery, [forceNumericSort])

sortQuery

Object containing desired ordering

forceNumericSort

Cast strings to number when sorting.

Examples

Advanced Select: Match a Field Without Specifying Array Index

• Mongo Docs

With MongoDB, you can search a document with a subarray of objects that you want to match when any one of the elements in the array matches. This tool implements it in SQL using a subquery, so it will likely not be the efficient on large datasets.

To enable subfield matching, you can pass a third parameter which is either an array of dotted paths that will be assumed to potentially be arrays or true if you want it to assume any field can be an array.

Example document:

{
  "courses": [{
      "distance": "5K"
    }, {
      "distance": "10K"
    }]
]

Example queries to match:

mongoToPostgres('data', { 'courses.distance': '5K' }, ['courses'])
mongoToPostgres('data', { 'courses.distance': '5K' }, true)

This then creates a PostgreSQL query like the following:

(data->'courses'->>'distance'='5K'
OR EXISTS (SELECT * FROM jsonb_array_elements(data->'courses')
           WHERE jsonb_typeof(data->'courses')='array' AND value->>'distance'='5K'))

Note: nested paths are not yet supported, so passing ['courses', 'courses.distance'] won't support checking both. The first matching path is the one that will be used.

Supported Features

• $eq, $gt, $gte, $lt, $lte, $ne
• $or, $not, $in, $nin
• $elemMatch
• $regex, $type, $size, $exists, $mod, $all

Todo

• Filtering
  • $expr
  • Bitwise Operators
• Update
  • $pop
  • $currentDate
  • $setOnInsert

Cannot Support

• $where

See also

• PostgreSQL json/jsonb functions and operators
• PostgreSQL json documentation
• MongoDB query documention
• PostgreSQL Array Functions
• JSON array to PostgreSQL Array