Convert query parameters from API urls to MongoDB queries (advanced querying, filtering, sorting, …)
- Powerful. Supports most of MongoDB operators (
$in
,$regexp
, …) and features (nested objects, projection, type casting, …) - Custom. Allows customization of keys (ie,
fields
vsselect
) and options - Agnostic. Works with any web frameworks (Express, Koa, …) and/or MongoDB libraries (mongoose, mongoskin, …)
- Simple. ~200 LOCs, dependency-free ES6 code
- Fully tested. 100% code coverage
npm i --save api-query-params
aqp(queryString, [opts])
Converts
queryString
into a MongoDB query object
queryString
: query string part of the requested API URL (ie,firstName=John&limit=10
). Works with already parsed object too (ie,{status: 'success'}
) [required]opts
: object for advanced options (See below) [optional]
The resulting object contains the following properties:
filter
which contains the query criteriaprojection
which contains the query projectionsort
,skip
,limit
which contains the cursor modifierspopulation
which contains the query population (mongoose feature only)
import aqp from 'api-query-params';
const query = aqp(
'status=sent×tamp>2016-01-01&author.firstName=/john/i&limit=100&skip=50&sort=-timestamp&populate=logs&fields=id,logs.ip'
);
// {
// filter: {
// status: 'sent',
// timestamp: { $gt: Fri Jan 01 2016 01:00:00 GMT+0100 (CET) },
// 'author.firstName': /john/i
// },
// sort: { timestamp: -1 },
// skip: 50,
// limit: 100,
// projection: { id: 1 },
// population: [ { path: 'logs', select: { ip: 1 } } ]
// }
import express from 'express';
import aqp from 'api-query-params';
import User from './models/User';
const app = express();
app.get('/users', (req, res, next) => {
const { filter, skip, limit, sort, projection, population } = aqp(req.query);
User.find(filter)
.skip(skip)
.limit(limit)
.sort(sort)
.select(projection)
.populate(population)
.exec((err, users) => {
if (err) {
return next(err);
}
res.send(users);
});
});
That's it. Your /users
endpoint can now query, filter, sort your User
mongoose model and more.
MongoDB | URI | Example | Result |
---|---|---|---|
$eq |
key=val |
type=public |
{filter: {type: 'public'}} |
$gt |
key>val |
count>5 |
{filter: {count: {$gt: 5}}} |
$gte |
key>=val |
rating>=9.5 |
{filter: {rating: {$gte: 9.5}}} |
$lt |
key<val |
createdAt<2016-01-01 |
{filter: {createdAt: {$lt: Fri Jan 01 2016 01:00:00 GMT+0100 (CET)}}} |
$lte |
key<=val |
score<=-5 |
{filter: {score: {$lte: -5}}} |
$ne |
key!=val |
status!=success |
{filter: {status: {$ne: 'success'}}} |
$in |
key=val1,val2 |
country=GB,US |
{filter: {country: {$in: ['GB', 'US']}}} |
$nin |
key!=val1,val2 |
lang!=fr,en |
{filter: {lang: {$nin: ['fr', 'en']}}} |
$exists |
key |
phone |
{filter: {phone: {$exists: true}}} |
$exists |
!key |
!email |
{filter: {email: {$exists: false}}} |
$regex |
key=/value/<opts> |
email=/@gmail\.com$/i |
{filter: {email: /@gmail.com$/i}} |
$regex |
key!=/value/<opts> |
phone!=/^06/ |
{filter: {phone: { $not: /^06/}}} |
For more advanced usage ($or
, $type
, $elemMatch
, etc.), pass any MongoDB query filter object as JSON string in the filter
query parameter, ie:
aqp('filter={"$or":[{"key1":"value1"},{"key2":"value2"}]}');
// {
// filter: {
// $or: [
// { key1: 'value1' },
// { key2: 'value2' }
// ]
// },
// }
- Useful to limit the number of records returned.
- Default operator keys are
skip
andlimit
.
aqp('skip=5&limit=10');
// {
// skip: 5,
// limit: 10
// }
- Useful to limit fields to return in each records.
- Default operator key is
fields
. - It accepts a comma-separated list of fields. Default behavior is to specify fields to return. Use
-
prefixes to return all fields except some specific fields. - Due to a MongoDB limitation, you cannot combine inclusion and exclusion semantics in a single projection with the exception of the _id field.
- It also accepts JSON string to use more powerful projection operators (
$
,$elemMatch
or$slice
)
aqp('fields=id,url');
// {
// projection: { id: 1, url: 1}
// }
aqp('fields=-_id,-email');
// {
// projection: { _id: 0, email: 0 }
// }
aqp('fields={"comments":{"$slice":[20,10]}}');
// {
// projection: { comments: { $slice: [ 20, 10 ] } }
// }
- Useful to sort returned records.
- Default operator key is
sort
. - It accepts a comma-separated list of fields. Default behavior is to sort in ascending order. Use
-
prefixes to sort in descending order.
aqp('sort=-points,createdAt');
// {
// sort: { points: -1, createdAt: 1 }
// }
- Useful to populate (reference documents in other collections) returned records. This is a mongoose-only feature.
- Default operator key is
populate
. - It accepts a comma-separated list of fields.
- It extracts projection on populated documents from the
projection
object.
aqp('populate=a,b&fields=foo,bar,a.baz');
// {
// population: [ { path: 'a', select: { baz: 1 } } ],
// projection: { foo: 1, bar: 1 },
// }
Any operators which process a list of fields ($in
, $nin
, sort and projection) can accept a comma-separated string or multiple pairs of key/value:
country=GB,US
is equivalent tocountry=GB&country=US
sort=-createdAt,lastName
is equivalent tosort=-createdAt&sort=lastName
Any operators can be applied on deep properties using .
notation:
aqp('followers[0].id=123&sort=-metadata.created_at');
// {
// filter: {
// 'followers[0].id': 123,
// },
// sort: { 'metadata.created_at': -1 }
// }
The following types are automatically casted: Number
, RegExp
, Date
and Boolean
. null
string is also casted:
aqp('date=2016-01-01&boolean=true&integer=10®exp=/foobar/i&null=null');
// {
// filter: {
// date: Fri Jan 01 2016 01:00:00 GMT+0100 (CET),
// boolean: true,
// integer: 10,
// regexp: /foobar/i,
// null: null
// }
// }
If you need to disable or force type casting, you can wrap the values with string()
, date()
built-in casters or by specifying your own custom functions (See below):
aqp('key1=string(10)&key2=date(2016)&key3=string(null)');
// {
// filter: {
// key1: '10',
// key2: Fri Jan 01 2016 01:00:00 GMT+0100 (CET),
// key3: 'null'
// }
// }
The following options are useful to change the operator default keys:
skipKey
: custom skip operator key (default isskip
)limitKey
: custom limit operator key (default islimit
)projectionKey
: custom projection operator key (default isfields
)sortKey
: custom sort operator key (default issort
)filterKey
: custom filter operator key (default isfilter
)populationKey
: custom populate operator key (default ispopulate
)
aqp('organizationId=123&offset=10&max=125', {
limitKey: 'max',
skipKey: 'offset',
});
// {
// filter: {
// organizationId: 123,
// },
// skip: 10,
// limit: 125
// }
The following options are useful to specify which keys to use in the filter
object. (ie, avoid that authentication parameter like apiKey
ends up in a mongoDB query). All operator keys are (sort
, limit
, etc.) already ignored.
blacklist
: filter on all keys except the ones specifiedwhitelist
: filter only on the keys specified
aqp('id=e9117e5c-c405-489b-9c12-d9f398c7a112&apiKey=foobar', {
blacklist: ['apiKey'],
});
// {
// filter: {
// id: 'e9117e5c-c405-489b-9c12-d9f398c7a112',
// }
// }
You can specify you own casting functions to apply to query parameter values, either by explicitly wrapping the value in URL with your custom function name (See example below) or by implictly mapping a key to a function (See Specify casting per param keys
below). Note that you can also override built-in casting functions: boolean
, date
,null
,number
,regex
and string
.
casters
: object to specify custom casters, key is the caster name, and value is a function which is passed the query parameter value as parameter.
aqp('key1=lowercase(VALUE)&key2=int(10.5)&key3=true', {
casters: {
lowercase: val => val.toLowerCase(),
int: val => parseInt(val, 10),
boolean: val => (val === 'true' ? '1' : '0'),
},
});
// {
// filter: {
// key1: 'value',
// key2: 10,
// key3: '1',
// }
// }
You can specify how query parameter values are casted by passing an object.
castParams
: object which map keys to casters (built-in or custom ones using thecasters
option).
aqp('key1=VALUE&key2=10.5&key3=20&key4=foo', {
casters: {
lowercase: val => val.toLowerCase(),
int: val => parseInt(val, 10),
},
castParams: {
key1: 'lowercase',
key2: 'int',
key3: 'string',
key4: 'unknown',
},
});
// {
// filter: {
// key1: 'value',
// key2: 10,
// key3: '20',
// key4: 'foo',
// }
// }
MIT © Loris Guignard