What's New in Pgx V5

This is the third part of the pgx v5 series. Let’s discuss the new features from the title. CollectRows and RowToStructByName # CollectRows allow us to fetch an array of rows and set them into a slice of structs. Further, RowToStructByName is a long awaited feature which allows to store a row into a struct. sqlx module became very popular in Go community exactly for this reason, but pgx now supports something similar. RowToStructByName is a generic function of the func(row CollectableRow) (T, error) and it’s not the only one of this type, there are also RowTo, RowToStructByPos and RowToStructByNameLax. Under the hood, RowToStructByName is using reflection so it may be slightly slower than classic way of scanning the rows, but if you don’t mind this, it is very easy to use. All field names are going to be mapped to lowercased public struct field names and you may also use db struct tags if some field name differs from the struct field name. This explanation may sound difficult, but the following example will actually show that it is actually very easy to use.

In this second part of the pgx v5 series, we are going to discuss the QueryRewriter interface and NamedArgs. Let’s first check out the signature of the Exec method of the pgxpool.Pool struct.

Go has very rich standard library which also contains database/sql module with generic interface over SQL databases. Further, lib/pq PostgreSQL driver is fully compatible with database/sql and provides all basic database operations. However, since this is generic implementation over all supported databases, using it is a trade-off not supporting some PostgreSQL specific features. Fans of PostgreSQL (in further text pg) may want more and here comes pgx in help. pgx is a very rich pg driver supporting LISTEN / NOTIFY, COPY, type mapping between pg and Go, all specific pg types, wire protocol and much more.