Repository navigation
Ignore mappings by default #1392
Description
As Augusto Breno described at Google Groups (https://groups.google.com/forum/?fromgroups#!searchin/mapstruct-users/ignore%7Csort:date/mapstruct-users/BDaK45b3rAg/V1amaHESBwAJ) it would be great to have the possibility of ignore mappings by default.
And yes, as @agudian said at the same topic, it can result a little odd based on the main purpose of MapStruct, but it can be very useful in cases like I'll try to describe:
In some cases you have a class that has copied values from another for historical purposes, for example, an invoice and a customer. And at the same time, both classes have some fields that have the same name.
class Invoice {
private Long id;
private String code;
// Customer historical fields
private String customerNoun;
private String customerSurname;
// getters and setters ....
}
class Customer {
private Long id;
private String code;
private String noun;
private String surname;
// getters and setters ....
}
Then, you may have following mapping definition to copy values from Customer to Invoice:
// ...
@Mappings({
@Mapping(source = "noun", target = "customerNoun"),
@Mapping(source = "surname", target = "customerSurname"),
@Mapping(target = "id", ignore = true),
@Mapping(target = "code", ignore = true)
})
public abstract void fillInvoice(Customer customer, @MappingTarget Invoice invoice);
// ...
If you have just two properties to ignore, it's not a big deal, but in more complex structs, it can be a large number of properties that match it's name, so it would be great to have something as Augusto described in the @Mapper annotation, or even in the @Mappings annotation such as:
// ...
@Mapper(ignoreByDefault=true)
// ...
or...
// ...
@Mappings(ignoreByDefault=true, {
@Mapping(source = "noun", target = "customerNoun"),
// ...
Regards!
Activity
sjaakd commentedon Mar 10, 2018
@mowcixo if you want something like this, I would prefer the
@BeanMappingiso@Mappings. In java8@Mappingsis marked by@Mappingas@Repeatable.@BeanMappingis there as holder specific for bean mapping methods.One solution might be: you could separate the ignore mappings into a specific mapper, lets say for the sake of the example:
LenientMapper. The main mapper (MainMapper) could include a uses relation to theLenientMapper. TheLenientMapperwould have the following mapper definition:@Mapper( unmappedTargetPolicy = ReportingPolicy.IGNORE). You could even work with object graphs like this:So:
WDYT?
sjaakd commentedon Mar 17, 2018
@mowcixo : I added an example based on the approach above. (see here). However, you would still have to ignore name based mappings explicitly.
However: would this approach offer a solution for you?
mowcixo commentedon Mar 19, 2018
@sjaakd thanks for the examples!
So, thinking in my case, I would need an intermediate class that only have the fields I need, and use it to fill the final class? I think it could work.
sjaakd commentedon Mar 19, 2018
Let me know. As the example points out: there's a difference between 'ignore' and 'not warn" for unmapped target properties. If you have implicit mappings (so source and target have the same name), MapStruct will map them and you'll still have to ignore them manually.
I'm working on a PR to do a complete ignore. But that's work-in-progress.
mapstruct#1392 add option to default ignoreAll mappings in a bean map…
mapstruct#1392 add option to default ignoreAll mappings in a bean map…
mapstruct#1392 add option to default ignoreAll mappings in a bean map…
mapstruct#1392 more meaningful test and small bugfix
mapstruct#1392 add option to default ignoreAll mappings in a bean map…
mapstruct#1392 add option to default ignoreAll mappings in a bean map…
mapstruct#1392 add option to default ignoreAll mappings in a bean map…
sjaakd commentedon Apr 21, 2018
implemented..
mowcixo commentedon Apr 23, 2018
@sjaakd Thanks!