Having trouble with parsing one single field from JSON response after enabling minify, with minify disabled all works correctly:
retrofit API call:
#FormUrlEncoded
#POST("/api/test")
fun test(#Field <some String fields>
): Observable<Response<TestListing>>
wrapped in repo
override fun test(<some String fields>): Observable<Response<TestListing>> {
return api.test(<some String fields>)
.subscribeOn(schedulers.io())
}
model:
data class TestListing (
#Json(name = "success") val success:Int,
#Json(name = "user") val user: TestUser?
)
TestUser class
data class TestUser(
#Json(name = "id") val id: Int,
#Json(name = "email") val email: String,
#Json(name = "name") val name: String,
#Json(name = "key") val remix_userkey: String,
#Json(name = "downloads_limit") val downloads_limit: Int?,
<some other fields>
)
and finally calling it in a viewModel
fun test(<some String fields>){
compositeDisposable.add(testRepo.test(<some String variables>)
.subscribeOn(schedulers.io())
.observeOn(schedulers.main())
.subscribe ({ testList ->
testListDebug.postValue(testList)
if (testList.isSuccessful) user.postValue(userList.body()?.user)
else {<some error posting>}
})
{ throwable -> <some actions>})
}
So without minifyEnabled it parses this JSON
{"success":1,"user":{"id":"123456","email":"test#test.com","name":"Test","remix_userkey":"abcd123abcd","downloads_limit":15}}
correctly, after I enable minify - id field is always 0.
Same JSON, but somehow it wraps in retrofit Response already with id=0 in the body(all other fields are parsed correctly)
example of testListDebug value from debugger after API call
Tried adding all library rules in proguard-rules.pro file, but with no effect; also tried adding #Keep annotation to TestUser class and renaming id field
Where I can dig from here? Is it something regarding Moshi or Retrofit/Okhttp?
Figured it out - needed to keep a custom moshi annotation class, which was used for parsing some field(which sometimes Int and sometimes Boolean) in other API calls and which was not used here. After adding keep annotation to it id is parsed fine
Very strange behavior since this annotation was not used here
Consider the following data model where status is provided by the API at the top level and message is provided at a nested level (this isn't really the case but just to illustrate the problem):
data class (
#Json(name = "status")
val status: Int,
#Transient
val message: String
)
Now, suppose I want to save this data model into my database. I'd change the model like so:
data class (
#ColumnInfo(name = "status")
#Json(name = "status")
val status: Int,
#ColumnInfo(name = "message")
#Transient
val message: String
)
the message field will not be saved into the database due to the Transient annotation. But what if I want to? How can I force Room to save the field but Moshi to ignore it.
The problem is I have no control over this API, or how it would change in the future, so removing #Transient and changing the name to some random complicated string is not an ideal solution, even if it will fix the problem.
Found a couple of related issues on Moshi:
Moshi/94
Moshi/1262
Apparently this is an age-old issue...
Maybe, can you use a mapper between your object and Room database ?
Like:
data class MessageEntity(
#ColumnInfo(name = "status")
val status: Int,
#ColumnInfo(name = "message")
val message: String
) {
// ...
}
data class Message (
#Json(name = "status")
val status: Int,
#Transient
val message: String
) {
fun toEntity(): MessageEntity {// ...}
fun fromEntity(e: MessageEntity): Message {// ...} // or Extensions
}
I'm trying to read a json response from a webservice, but without success.
This is the json i receive:
{
"rsp": {
"#code": "0",
"#message": ""
},
"listOfStrings":[]
}
And this is relative data class where i parse response
data class Response(
val rsp : Rsp,
val listOfStrings : List<String>
)
data class Rsp(
#Json(name = "#code")
val code : String,
#Json(name = "#message")
val message : String
)
But it seems that moshi for some reason it's not able to parse json into object, because i always get Response object with all null fields.
So what's wrong? May the "#" character of json response fields cause problems?
UPDATE
Now i can parse correctly response by change #Json annotation into #field:Json:
data class Rsp(
#field:Json(name = "#code")
val code : String,
#field:Json(name = "#message")
val message : String
)
But i'm curious to know why it works.
#field:Json is required if you want moshi-kotlin to work with proguard according to the discussion here: https://github.com/square/moshi/issues/315
Try this model and let me know if it works:
#Parcelize
data class Response(
#Json(name = "rsp")
val rsp: Rsp,
#Json(name = "listOfStrings")
val listOfStrings: List<String>
) : Parcelable {
#Parcelize
data class Rsp(
#Json(name = "#code")
val code: String,
#Json(name = "#message")
val message: String
) : Parcelable
}
Edit:
If it didn't work, try to add back-slash behind those field names that have #.
Like: #Json(name = "\#code").
UPDATE AFTER QUESTION GOT UPDATE:
You need to add moshi-kotlin dependency and then using KotlinJsonAdapterFactory
val moshi = Moshi.Builder()
// ... add your own JsonAdapters and factories ...
.add(KotlinJsonAdapterFactory())
.build()
Then moshi couldn't ignore #Json.
I'm integrating with the Room persistence library. I have a data class in Kotlin like:
#Entity(tableName = "story")
data class Story (
#PrimaryKey val id: Long,
val by: String,
val descendants: Int,
val score: Int,
val time: Long,
val title: String,
val type: String,
val url: String
)
The #Entity and #PrimaryKey annotations are for the Room library. When I try to build, it is failing with error:
Error:Cannot find setter for field.
Error:Execution failed for task ':app:compileDebugJavaWithJavac'.
> Compilation failed; see the compiler error output for details.
I also tried providing a default constructor:
#Entity(tableName = "story")
data class Story (
#PrimaryKey val id: Long,
val by: String,
val descendants: Int,
val score: Int,
val time: Long,
val title: String,
val type: String,
val url: String
) {
constructor() : this(0, "", 0, 0, 0, "", "", "")
}
But this doesn't work as well. A thing to note is that it works if I convert this Kotlin class into a Java class with getters and setters. Any help is appreciated!
Since your fields are marked with val, they are effectively final and don't have setter fields.
Try switching out the val with var.
You might also need to initialize the fields.
#Entity(tableName = "story")
data class Story (
#PrimaryKey var id: Long? = null,
var by: String = "",
var descendants: Int = 0,
var score: Int = 0,
var time: Long = 0L,
var title: String = "",
var type: String = "",
var url: String = ""
)
EDIT
The above solution is a general fix for this error in Kotlin when using Kotlin with other Java libraries like Hibernate where i've seen this as well. If you want to keep immutability with Room, see some of the other answers which may be more specific to your case.
In some cases immutability with Java libraries is simply not working at all and while making sad developer noises, you have to switch that val for a var unfortunately.
Hey I don't know if everyone know or not, but you can not have column which is starting from is into Room.
For example you can't have like this
#Entity(tableName = "user")
data class User (
#PrimaryKey var id: Long? = null,
var userName: String = "",
var isConnectedToFB: Boolean = false,
)
If you have #Ignore field in the data class constructor you need to move it to class body like this:
#Entity(primaryKeys = ["id"])
data class User(
#field:SerializedName("id")
val id: Int,
#field:SerializedName("name")
val name: String,
#field:SerializedName("age")
val age: Int
) {
#Ignore
val testme: String?
}
All kudos go to marianperca on GitHub: https://github.com/android/architecture-components-samples/issues/421#issuecomment-442763610
There is an issue in room db library java code generation.
I was using optional field isFavorite. It gives me same error then I change my field name to favorite then compiled.
before
var isFavorite: Int? = 0,
after changing working fine
var favorite: Int? = 0,
Thanks
According to https://stackoverflow.com/a/46753804/2914140 if you have an autogenerated primary key, you should write so:
#Entity(tableName = "story")
data class Story (
val by: String,
val descendants: Int,
val score: Int,
val time: Long,
val title: String,
val type: String,
val url: String
) {
#PrimaryKey(autoGenerate = true)
var id: Int = 0
}
Note that #PrimaryKey is written inside the class body and contains modifier var.
If you later want to update a row in a database with different parameters, use these lines:
val newStory = story.copy(by = "new author", title = "new title") // Cannot use "id" in object cloning
newStory.id = story.id
dao.update(newStory)
UPDATE
I still don't use AndroidX, and Room is 'android.arch.persistence.room:runtime:1.1.1'.
You can extend this class from Serializable. But if you want to extend it from Parcelable, you will get a warning (over id variable): Property would not be serialized inro a 'Parcel'. Add '#IgnoredOnParcel' annotation to remove this warning:
Then I moved an id from the body to the constructor. In Kotlin I use #Parcelize to create Parcelable classes:
#Parcelize
#Entity(tableName = "story")
data class Story (
#PrimaryKey(autoGenerate = true)
var id: Int = 0,
val by: String,
val descendants: Int,
val score: Int,
val time: Long,
val title: String,
val type: String,
val url: String
) : Parcelable
Had this error in Java.
You cannot have a column starting with is or is_ in Java.
Try renaming the column.
Another solution:
You either have to pass the field in the constructor and initialize it with the constructor argument, or create a setter for it.
Example:
public MyEntity(String name, ...) {
this.name = name;
...
}
public void setName(String name) {
this.name = name;
}
This error will be thrown if your column starts with Is:
#ColumnInfo(name = "IsHandicapLeague")
#NonNull
var isHandicapLeague: String = "Y"
Add a default set() function to eliminate
fun setIsHandicapLeague(flag:String) {
isHandicapLeague = flag
}
Just make the variables mutable, change val into var for Kotlin, Or private into public for Java
This is a bug and is fixed in Room 2.1.0-alpha01
https://developer.android.com/jetpack/docs/release-notes#october_8_2018
Bug Fixes
Room will now properly use Kotlin’s primary constructor in
data classes avoiding the need to declare the fields as vars.
b/105769985
I've found that another cause of this compilation error can be due to the use of the Room's #Ignore annotation on fields of your entity data class:
#Entity(tableName = "foo")
data class Foo(
// Okay
#PrimaryKey
val id: String,
// Okay
val bar: String,
// Annotation causes compilation error, all fields of data class report
// the "Cannot find setter for field" error when Ignore is present
#Ignore
val causeserror: String
)
The same error also seems to happens when using the #Transient annotation.
I've noticed this issue using version 2.2.2 of Room:
// build.gradle file
dependencies {
...
kapt "androidx.room:room-compiler:2.2.2"
...
}
Hope that helps someone!
You can try to rename id variable to another name. It worked for me ;
var id: Long? = null
to
var workerId: Long? = null
If you have to name as id and you are using retrofit, then you may need to add SerializedName("id")
Another cause of this may be the naming of the field. If you use any of the pre-defined keywords, you will get the same error.
For instance, you can not name your column "is_active".
Reference: http://www.sqlite.org/lang_keywords.html
It seems like Room and Kotlin versions need to be matched. I have same issue with Room 2.3.0 and Kotlin 1.6.10 but it's ok with Kotlin 1.5.20. It looks ok after I updated Room to 2.4.2.
https://youtrack.jetbrains.com/issue/KT-45883
Also there is a possible solution to use #JvmOverloads constructor for better Java compability.
Updating Room library to the latest version 2.4.2 solve the issue
The correct way to fix this issue would be simply updating to Room v2.4.3 or higher.
Workaround
If you're running on an older version of Room, one that uses an old version of the kotlinx-metadata-jvm library which doesn't understand 1.5.x metadata, a simple workaround would be adding the following line to your build.gradle:
kapt "org.jetbrains.kotlinx:kotlinx-metadata-jvm:0.5.0"
Source: https://youtrack.jetbrains.com/issue/KT-45883/KAPT-Cannot-find-setter-for-field-compiling-projects-with-Room-db-breaks-using-150-M2
Just an update if somebody comes across this thread in 2019, after spending hours digging online on why this should work, but it doesn't.
Using val works as expected if you are using the AndroidX version ( androidx.room:room-<any>:2.*) but it doesn't when using the old android.arch.persistence.room:<any>:1.1.1 and it seems that version 2.* wasn't released on this latter repo.
Edit: typos
If you want the val immutability available for your entity, it is possible.
You should update to AndroidX room current version.
Check for the related issue here it is marked as Won't Fix
Now they have release a fix related to the issue with version 2.0.0-beta01
Now you can use immutable val with default value e.g:
#Entity("tbl_abc")
data class Abc(
#PrimaryKey
val id: Int = 0,
val isFavourite: Boolean = false
)
Previously, the above snippet will throw an error of Cannot find setter for field. Changing into var is a great workaround, but I prefer for the entity class to be immutable from outside invocation
You can now start your field with is but you can't have a number next to the is like : is2FooSelected, you have to rename to isTwoFooSelected.
I think that the variable we wrote as id is getting mixed up with the id in the system. Therefore, when I define it as uuid, my error is resolved. I think it will be solved too. Also, try using var instead of val.
#PrimaryKey(autoGenerate = true)
var uuid:Int=0
Just use var instead of val and if you are using private keyword, make it public.
#Entity(tableName = "story")
data class Story (
#PrimaryKey val id: Long,
var by: String,
var descendants: Int,
var score: Int,
var time: Long,
var title: String,
var type: String,
var url: String
)