pg.honey — com.github.igrishaev/pg2-honey 0.1.4

HoneySQL wrappers and shortcuts.

delete clj

(delete conn table)

(delete conn table {:as opt :keys [where returning] :or {returning [:*]}})

execute clj

(execute conn sql-map)

(execute conn sql-map {:as opt :keys [honey]})

Like pg.core/execute but accepts a HoneySQL map which gets rendered into SQL vector and split on a query and parameters.

Arguments:

• conn: a Connection object;
• sql-map: a map like {:select [:this :that] :from [...]}
• opt: PG2 options; pass the :honey key for HoneySQL params.

Result:

• same as pg.core/execute.

find clj

(find conn table kv)

(find conn
      table
      kv
      {:as opt :keys [fields limit offset order-by] :or {fields [:*]}})

Select rows by a column->value filter, for example:

{:name 'Foo', :active true}

All the kv pairs get concatenated with AND.

The optional arguments are:

• fields (default is [:*])
• limit
• offset
• order-by

find-first clj

(find-first conn table kv)

(find-first conn
            table
            kv
            {:as opt :keys [fields offset order-by] :or {fields [:*]}})

Like find but gets the first row only. Adds the limit 1 expression to the query. Supports the same optional arguments.

format clj

(format sql-map)

(format sql-map opt)

Like honey.sql/format but with some Postgres-related overrides. Namely, it produces a SQL expressions with dollar signs instead of question marks, e.g. SELECT * FROM USERS WHERE id = $1.

get-by-id clj

(get-by-id conn table id)

(get-by-id conn table id {:as opt :keys [pk fields] :or {pk :id fields [:*]}})

Get a single row by its primary key.

By default, the name of the primary key is :id which can be overridden by the :pk parameter.

The optimal :fields vector specifies which columns to return, [:*] by default.

get-by-ids clj

(get-by-ids conn table ids)

(get-by-ids conn
            table
            ids
            {:as opt :keys [pk fields order-by] :or {pk :id fields [:*]}})

Get multiple rows from a table by their PKs. It's not recommended to pass thousands of PKs at once. Split them by smaller chunks instead.

Like get-by-id, accepts the PK name, the column names to return, and order-by clause.

Returns a vector of rows.

HONEY_OVERRIDES clj

insert clj

(insert conn table kvs)

(insert conn
        table
        kvs
        {:as opt
         :keys [returning on-conflict do-update-set do-nothing]
         :or {returning [:*]}})

Insert a collection of rows into a table. The kvs argument must be a seq of maps.

The optional arguments are:

• returning: a vector of columns to return, default is [:*]
• on-conflict,
• do-update-set,
• do-nothing: Postgres-specific ON CONFLICT ... DO ... expressions.

Other PG2-related options are supported.

The result depends on the returning clause and PG2 options. By default, it will be a vector of inserted rows.

insert-one clj

(insert-one conn table kv)

(insert-one conn
            table
            kv
            {:as opt
             :keys [returning on-conflict do-update-set do-nothing]
             :or {returning [:*]}})

Like insert but accepts a single row.

Supports the same options. The default result is a single inserted row.

prepare clj

(prepare conn sql-map)

(prepare conn sql-map {:as opt :keys [honey]})

Prepare a statement expressed wit a Honey map. For parameters, use either blank values or raw expressions, for example:

{:select [:*] :from :users :where [:= :id 0]}

or

{:select [:*] :from :users :where [:raw 'id = $1']}

Return an instance of PreparedStatement class.

query clj

(query conn sql-map)

(query conn sql-map {:as opt :keys [honey]})

Like pg.core/query but accepts a HoneySQL map which gets rendered into a SQL string.

Arguments:

• conn: a Connection object;
• sql-map: a map like {:select [:this :that] :from [...]}
• opt: PG2 options; pass the :honey key for HoneySQL params.

Result:

• the same as pg.core/query.

update clj

(update conn table kv)

(update conn table kv {:as opt :keys [where returning] :or {returning [:*]}})

Update rows in a table. The kv is a map of field->value. Note that the value might be not a scalar but something like [:raw 'now()'] or any other HoneySQL expression.

The optional parametes are:

• where which acts like a filter when updating the rows;
• returning which determines what columns to return.