Usually, we need to write a lot of boilerplate code that builds dynamic queries through if statements. Conventional code generation methods can only generate equality conditions. I thought of solving this problem by adding a predicate suffix to fields, and created the tool gooogen. The usage is as follows:
//go:generate gooogen
type UserQuery struct {
PageQuery
Name *string
NameIn *[]string
ScoreGe *int
ScoreLt *int
}
Run go generate will generate the following code:
func (q UserQuery) BuildConditions() ([]string, []any) {
conditions := make([]string, 0, 4)
args := make([]any, 0, 4)
if q.Name != nil {
conditions = append(conditions, "name = ?")
args = append(args, *q.Name)
}
if q.NameIn != nil {
phs := make([]string, 0, len(*q.NameIn))
for _, arg := range *q.NameIn {
args = append(args, arg)
phs = append(phs, "?")
}
conditions = append(conditions, "name IN ("+strings.Join(phs, ", ")+")")
}
if q.ScoreGe != nil {
conditions = append(conditions, "score >= ?")
args = append(args, *q.ScoreGe)
}
if q.ScoreLt != nil {
conditions = append(conditions, "score < ?")
args = append(args, *q.ScoreLt)
}
return conditions, args
}
Combining it with entity objects in the ORM can build complete query statements, thereby simplifying the development of dynamic queries.
Feedback is welcome.