Upsert in KnexJS - sql

Question

I have an upsert query in PostgreSQL like:
INSERT INTO table
(id, name)
values
(1, 'Gabbar')
ON CONFLICT (id) DO UPDATE SET
name = 'Gabbar'
WHERE
table.id = 1
I need to use knex to this upsert query. How to go about this?

Answer 1

So I solved this using the following suggestion from Dotnil's answer on Knex Issues Page:
var data = {id: 1, name: 'Gabbar'};
var insert = knex('table').insert(data);
var dataClone = {id: 1, name: 'Gabbar'};
delete dataClone.id;
var update = knex('table').update(dataClone).whereRaw('table.id = ' + data.id);
var query = `${ insert.toString() } ON CONFLICT (id) DO UPDATE SET ${ update.toString().replace(/^update\s.*\sset\s/i, '') }`;
return knex.raw(query)
.then(function(dbRes){
// stuff
});
Hope this helps someone.

Answer 2

As of knex#v0.21.10+ a new method onConflict was introduced.
Official documentation says:
Implemented for the PostgreSQL, MySQL, and SQLite databases. A
modifier for insert queries that specifies alternative behaviour in
the case of a conflict. A conflict occurs when a table has a PRIMARY
KEY or a UNIQUE index on a column (or a composite index on a set of
columns) and a row being inserted has the same value as a row which
already exists in the table in those column(s). The default behaviour
in case of conflict is to raise an error and abort the query. Using
this method you can change this behaviour to either silently ignore
the error by using .onConflict().ignore() or to update the existing
row with new data (perform an "UPSERT") by using
.onConflict().merge().
So in your case, the implementation would be:
knex('table')
.insert({
id: id,
name: name
})
.onConflict('id')
.merge()

Answer 3

I've created a function for doing this and described it on the knex github issues page (along with some of the gotchas for dealing with composite unique indices).
const upsert = (params) => {
const {table, object, constraint} = params;
const insert = knex(table).insert(object);
const update = knex.queryBuilder().update(object);
return knex.raw(`? ON CONFLICT ${constraint} DO ? returning *`, [insert, update]).get('rows').get(0);
};
Example usage:
const objToUpsert = {a:1, b:2, c:3}
upsert({
table: 'test',
object: objToUpsert,
constraint: '(a, b)',
})
A note about composite nullable indices
If you have a composite index (a,b) and b is nullable, then values (1, NULL) and (1, NULL) are considered mutually unique by Postgres (I don't get it either).

Answer 4

Yet another approach I could think of!
exports.upsert = (t, tableName, columnsToRetain, conflictOn) => {
const insert = knex(tableName)
.insert(t)
.toString();
const update = knex(tableName)
.update(t)
.toString();
const keepValues = columnsToRetain.map((c) => `"${c}"=${tableName}."${c}"`).join(',');
const conflictColumns = conflictOn.map((c) => `"${c.toString()}"`).join(',');
let insertOrUpdateQuery = `${insert} ON CONFLICT( ${conflictColumns}) DO ${update}`;
insertOrUpdateQuery = keepValues ? `${insertOrUpdateQuery}, ${keepValues}` : insertOrUpdateQuery;
insertOrUpdateQuery = insertOrUpdateQuery.replace(`update "${tableName}"`, 'update');
insertOrUpdateQuery = insertOrUpdateQuery.replace(`"${tableName}"`, tableName);
return Promise.resolve(knex.raw(insertOrUpdateQuery));
};

Answer 5

very simple.
Adding onto Dorad's answer, you can choose specific columns to upsert using merge keyword.
knex('table')
.insert({
id: id,
name: name
})
.onConflict('id')
.merge(['name']); // put column names inside an array which you want to merge.