sinopower / jet Goto Github PK

Jet is a super-flexible lightweight SQL interface (GoDoc)

Features:

• En/decode custom types using ComplexValue
• Prepared statement LRU cache
• Unpack query results to any format
• Can expand queries for map and slice arguments (e.g. $1 to $1, $2, $3, useful for hstore or set membership queries)
• Serializes hstore columns to maps
• Simple migration API
• Customizable column name mapper

db, err := jet.Open("postgres", "...")

db.Query(`INSERT INTO "fruits" ( "name", "price" ) VALUES ( $1, $2 )`, "banana", 2.99).Run()

Run is Jet's Exec equivalent and is used instead of Rows() when no return values are expected

var rows []*struct{
  Name  string
  Price int
}
db.Query(`SELECT * FROM "fruits"`).Rows(&rows)

Jet's column mapper is very powerful. It tries to map the columns to any value you provide. You're not required to use a fixed output format. In this case rows could be anything e.g struct, *struct, []struct, []*struct, Type, *Type, []Type, []*Type even map[string]interface{} or just simple values like int or *int. You get the idea.

Jet can also deserialize hstore columns for you. In this case the header column is a hstore value.

var out struct{
  Header  map[string]interface{}
  Body    string
}
db.Query(`SELECT * FROM "emails"`).Rows(&out)

If you want to do e.g. hstore inserts or set membership queries, Jet can automatically expand the query and adjust the argument list for you.

• Maps expand to k1, v1, k2, v2, ...
• Slices expand to v1, v2, ...

Passing in a map argument

db.Query(`INSERT INTO "emails" ( "header", "body" ) VALUES ( hstore(ARRAY[ $1 ]), $2 )`, aMap, aBody).Run()

will expanded the query to

INSERT INTO "emails" ( "header", "body" ) VALUES ( hstore(ARRAY[ $1, $2, $3, $4 ... ]), $5 )

Passing in a slice argument

db.Query(`SELECT * FROM "files" WHERE "files"."name" IN ( $1 )`, aSlice)

will expand the query to

SELECT * FROM "files" WHERE "files"."name" IN ( $1, $2, $3, ... )