(deftransforms model column->direction->fn)
Define type transforms to use for a specific model. transforms
should be a map of
{:column-name {:in <fn>, :out <fn>}}
:in
transforms are applied to values going over the wire to the database; these generally only applied to values
passed at or near the top level to various functions; don't expect Toucan 2 to parse your SQL to find out which
parameter corresponds to what in order to apply transforms or to apply transforms inside JOINS in hand-written
HoneySQL. That said, unless you're doing something weird your transforms should generally get applied.
:out
transforms are applied to values coming out of the database; since nothing weird really happens there this is
done consistently.
Transform functions for either case are skipped for nil
values.
Example:
(deftransforms :models/user
{:type {:in name, :out keyword}})
You can also define transforms independently, and derive a model from them:
(deftransforms ::type-keyword
{:type {:in name, :out keyword}})
(derive :models/user ::type-keyword)
(derive :models/user ::some-other-transform)
Don't derive a model from multiple deftransforms
for the same key in the same direction.
When multiple transforms match a given model they are combined into a single map of transforms with merge-with merge
. If multiple transforms match a given column in a given direction, only one of them will be used; you should
assume which one is used is indeterminate. (This may be made an error, or at least a warning, in the future.)
Until upstream issue https://github.com/camsaul/methodical/issues/97 is resolved, you will have to specify which method should be applied first in cases of ambiguity using [[methodical.core/prefer-method!]]:
(m/prefer-method! transforms ::user-with-location ::user-with-password)
If you want to override transforms completely for a model, and ignore transforms from ancestors of a model, you can
create an :around
method:
(defmethod toucan2.tools.transforms :around ::my-model
[_model]
{:field {:in name, :out keyword}})
Define type transforms to use for a specific model. `transforms` should be a map of ```clj {:column-name {:in <fn>, :out <fn>}} ``` `:in` transforms are applied to values going over the wire to the database; these generally only applied to values passed at or near the top level to various functions; don't expect Toucan 2 to parse your SQL to find out which parameter corresponds to what in order to apply transforms or to apply transforms inside JOINS in hand-written HoneySQL. That said, unless you're doing something weird your transforms should generally get applied. `:out` transforms are applied to values coming out of the database; since nothing weird really happens there this is done consistently. Transform functions for either case are skipped for `nil` values. Example: ```clj (deftransforms :models/user {:type {:in name, :out keyword}}) ``` You can also define transforms independently, and derive a model from them: ```clj (deftransforms ::type-keyword {:type {:in name, :out keyword}}) (derive :models/user ::type-keyword) (derive :models/user ::some-other-transform) ``` Don't derive a model from multiple [[deftransforms]] for the same key in the same direction. When multiple transforms match a given model they are combined into a single map of transforms with `merge-with merge`. If multiple transforms match a given column in a given direction, only one of them will be used; you should assume which one is used is indeterminate. (This may be made an error, or at least a warning, in the future.) Until upstream issue https://github.com/camsaul/methodical/issues/97 is resolved, you will have to specify which method should be applied first in cases of ambiguity using [[methodical.core/prefer-method!]]: ```clj (m/prefer-method! transforms ::user-with-location ::user-with-password) ``` If you want to override transforms completely for a model, and ignore transforms from ancestors of a model, you can create an `:around` method: ```clj (defmethod toucan2.tools.transforms :around ::my-model [_model] {:field {:in name, :out keyword}}) ```
(transforms model₁)
Return a map of
{column-name {:in <fn>, :out <fn>}}
For a given model
, all matching transforms are combined with merge-with merge
in an indeterminate order, so don't
try to specify multiple transforms for the same column in the same direction for a given model -- compose your
transform functions instead if you want to do that. See toucan2.tools.transformed/deftransforms
for more info.
Return a map of ```clj {column-name {:in <fn>, :out <fn>}} ``` For a given `model`, all matching transforms are combined with `merge-with merge` in an indeterminate order, so don't try to specify multiple transforms for the same column in the same direction for a given model -- compose your transform functions instead if you want to do that. See [[toucan2.tools.transformed/deftransforms]] for more info.
cljdoc is a website building & hosting documentation for Clojure/Script libraries
× close