Random question, Is there a way to stop `CacheAndNetwork` from emitting 2 values when the cache valu...

# apollo-kotlin

[Spencer]

Random question, Is there a way to stop CacheAndNetwork from emitting 2 values when the cache value and network value are the same?

[Stylianos Gakis]

If the items are equal, you could do .distinctUntilChanged on the flow itself. Don’t know if there’s something that can be done on the apollo-kotlin side though for this, or if something should be done there.

[Spencer]

That was a good idea, but when I tried it with

.distinctUntilChanged { prev, current ->
    prev.data?.equals(current.data)
}

it was always coming up false

[Stylianos Gakis]

So you said that your values are the same and you don’t want to re-emit. Yet the equals method returns false for them. So either
• They are not actually the same
• Their equals method is somehow broken
Have you checked why this returns false? Have you debugged and made sure those two objects are in fact the exact same data? Try something like that and if it still doesn’t work we can maybe figure out what’s going wrong.

Also as a side note, where are you applying this distinctUntilChanged? Showing some more code may help in case there’s some small mistake somewhere

[Spencer]

Yeah I have checked with the debugger and they are the exact same data.

Whats really weird is doing prev.data.toString() == current.data.toString() makes it return true then.

apolloClient.query(query)
    .fetchPolicy(CacheAndNetwork)
    .watch()
    .distinctUntilChanged { prev, current ->
        prev.data.toString() == current.data.toString()
    }
    .collect { response ->
        //....
    }
}.catch {
    //....
}

[Spencer]

My hope was that if we use CacheAndNetwork, and the response are the exact same, there really isn’t much of a need to rerun the collect again

[Stylianos Gakis]

But in the debugger, doing .equals also returns false I suppose right? Even though their toStrings() return true.
This may mean that their toString() is not really printing out everything, and that thing which is not printed out is different between them.
Other than that, I can’t imagine how else this could be failing. Remind me, are those generated classes all data classes, or do they have their equals methods manually generated somehow?

[Spencer]

Correct, even in the debugger doing the equal is false. I just checked that the toString length matches the length of the network response payload from the network profiler.

All of the generated classes are Data classes, but I don't see any equal() in the generated apollo types

[Spencer]

There must be something weird going on because the hash code is also different 😕

[Stylianos Gakis]

Can you try doing the same but with a smaller object? One that you can check with your own eyes to see if there's something different? This object looks way too big to check manually

[Stylianos Gakis]

Can it be something tricky like a list of items being returned with different ordering even though they're the same size or something silly like that? 😅

[Spencer]

All the time I was doing it, I was taking the raw data from the debugger and sticking it in a online diff checker

[Spencer]

🤦 Well I figured it out. I had a custom scalar but it was a class and not a data class. Turning it into a data class fixed it immediately. Thanks for all your help talking through it!!!

[Stylianos Gakis]

Heh so it was something to do with data classes 😄
You could also implement equals yourself properly instead of converting it into a data class if you don’t need that.
But so glad to help you, very happy that it now works 🎉