PostgreSQL

Learn how to build an app in Noloco around your PostgreSQL database

Overview

In this guide we are going to walkthrough how easy it is to connect your PostgreSQL database(s) to your Noloco app. Once connected, this will allow your team, customers &/or third party stakeholders to read, update and create records in your PostgreSQL instance directly from your Noloco app.

Connect your PostgreSQL database

  1. Add your PostgreSQL data source Navigate to the data tab in your Noloco app and click to add a new data source. From the list, choose Postgres.

2. Name your data source It's best practice to call the data source the same name as your database to help you keep track.

3. Enter the server connection information The hostname is the URL that you access your server on and the port is the port you use to connect. By default PostgreSQL uses port 5432 so if you're not sure try that.

4. Enter the database information We need the name of the database on the server that you want us to connect to and the schema within the database that we should pull the tables from. By default the schema is probably called public, however the database will be something that you named.

5. Enter the login details You will need to provide a PostgreSQL user's login details for us to use. Specifically we need a the username and password for a user with SELECT, INSERT, UPDATE and DELETE permissions for tables within the schema you are importing to Noloco.

Connecting as a read-only user

We support connecting to your database with only SELECT permissions, however you need to toggle this setting on in the last step of the new data source form.

Enabling SSL on connections

We support connecting to your database with SSL, just toggle it on in the last step of the new data source form.

Whitelisting Noloco's IP addresses

If you restrict connections to your PostgreSQL database by IP, you can whitelist our three static IP addresses that we might connect from.

18.203.60.136
54.217.27.248
54.228.83.124

Syncing tables

Noloco will import all tables from your PostgreSQL schema that have a primary key and whose name has some alpha-numeric or emoji characters. So for instance a table with just punctuation in the name would be ignored, as would a table with a valid name but no primary key.

We support tables with composite primary keys.

We only support regular tables and do not import views.

Built-in data types

We categorise built-in data types of columns into three buckets; fully supported, partially supported and unsupported. When importing a column from a table, you will only see it in Noloco if its type is fully or partially supported.

We do not support arrays of any of the supported types.

Fully supported column types

Data types that Noloco fully supports will be imported into your Noloco project with full read-write capabilities to be updated as well as displayed.

Postgres TypeNoloco Type

bigint

INTEGER

bigserial

INTEGER

bool

BOOLEAN

boolean

BOOLEAN

char

TEXT

character

TEXT

character varying

TEXT

date

DATE

decimal

DECIMAL

double precision

DECIMAL

float4

DECIMAL

float8

DECIMAL

int

INTEGER

int2

INTEGER

int4

INTEGER

int8

INTEGER

integer

INTEGER

interval

DURATION

money

DECIMAL (CURRENCY)

numeric

DECIMAL

real

DECIMAL

serial

INTEGER

serial2

INTEGER

serial4

INTEGER

serial8

INTEGER

smallint

INTEGER

smallserial

INTEGER

text

TEXT

timestamp

DATE

timestamp with timezone

DATE

timestamp without timezone

DATE

varchar

TEXT

Partially supported column types

Other data types are partially supported which means that we will import them to your Noloco project and display them, but we do not support any write operations to them so you may not update them from Noloco.

Postgres TypeNoloco Type

cidr

TEXT

inet

TEXT

json

TEXT

jsonb

TEXT

macaddr

TEXT

macaddr8

TEXT

pg_lsn

TEXT

uuid

TEXT

xml

TEXT

Unsupported column types

All other column types are unsupported and columns with those types will not be imported to Noloco (although the rest of the table's columns with supported types will be imported).

Custom data types

We support a limited set of custom data types created via a CREATE TYPE query. Specifically we will sync enumerated types as Noloco SINGLE_OPTION types. We will only sync these if the enum created is not empty. Right now we do not support enum arrays or any other custom types.

Domain types

We offer support for user-defined types created via a CREATE DOMAIN query. During schema syncing, domains will be mapped onto their underlying type and provided we offer full or partial support for that underlying type it will be synced to Noloco.

Foreign keys

Any table with a foreign key constraint on one or more of its columns to another synced table will have those foreign key(s) interpreted as a Noloco relationship field(s) when synced.

We do not support composite foreign keys. Only foreign key constraints on single columns will be synced as relationships in Noloco.

The type of relationship that will be created in Noloco depends on other constraints on the foreign key constrained column. However because PostgreSQL requires all columns referenced by a foreign key to have a unique constraint, they will always be one of the two relationships below.

FK-Constrained Column is UniqueNoloco Relationship

Yes

ONE_TO_ONE

No

MANY_TO_ONE

These relationships are fully functional Noloco relationships and any updates to the relationship values in Noloco will be propagated to the foreign key on your PostgreSQL database.

Join tables

As documented above, all PostgreSQL foreign keys must reference a column with a unique constraint. This means that from one foreign key alone you cannot create ONE_TO_MANY or MANY_TO_MANY relationships in PostgreSQL. The way that these are typically created are with join tables with multiple foreign keys that sit between two tables you want to relate. Noloco offers some limited support to interpret such join tables as multi-relationships.

Consider the following example:

Table A (id, value)
Table B (id, value)

Join Table (id, a_id, b_id)

The Join Table allows a MANY_TO_MANY relationship to be described between Table A and Table B by storing a normalised mapping of a_id and b_id.

If a table in your database matches all of these criteria, then we will classify it as a join table:

  • It has exactly two foreign keys (pointing to separate tables)

  • Either one or neither of the foreign keys has a unique constraint

  • All columns in the table are either in the primary key or in one of the foreign keys

Join tables are not synced as their own data type like the other tables. Instead they will be synced as a relationship field on each side of the relationship they point to. The type of Noloco relationship that is created for a join table depends on whether there are any unique constraints in the join table for the two foreign keys.

Number of FKs With Unique ConstraintsNoloco Relationship

0

MANY_TO_MANY

1

ONE_TO_MANY

2

N/A - not classified as join table

These relationships are fully functional Noloco relationships and any updates to the relationship values in Noloco will be propagated to the join table on your PostgreSQL database.

Reserved field names

The following (and any case variants) are reserved field names in Noloco:

  • createdAt

  • id

  • updatedAt

  • uuid

If your PostgreSQL table has an id column we will still import it to Noloco, however it will be renamed to [Table Name] Id to avoid conflicting with our own id.

If you have createdAt or updatedAt field(s) and they are of type date, timestamp, timestamp without timezone or timestamp with timezone, we will map your column(s) to our own field(s). Otherwise we will manage the created/updated times ourselves.

If you have a uuid column it will not be imported.

File fields

We support syncing columns containing a URL as a Noloco FILE field. Just open up the corresponding collection in your data table and click on the field then change its type to FILE. After the collection next has a data sync you will see your files at those URLs appearing in Noloco.

Create query collections

This will take you the the query editor where you can set up the query that will build your collection.

At the top of this page is an input for you to name your collection. On the left you can see an overview of this data source from Noloco and on the right is a query editor for you to develop and test your query over time. The results of testing your query will appear in the section at the bottom of the page. To save your query you must have a successful test of it.

Any errors in your query will be surfaced in the section at the bottom of the page.

After saving the query you will be able to see and use it like any other data type in Noloco. It will also be kept in sync with the upstream database both in schema and in data. If you want to edit the query at any time you can do so by clicking on the edit button from the data table.

Troubleshooting

Editing a relationship appears to create a new (empty) record

Double check the types of the columns in your foreign key constraint, if one is a bigint or int8 and another is an int or int4 we will be treating the values differently and may not build the correct relationship in Noloco. We would recommend to keep any relationship fields the exact same type as the field they reference.

Last updated