如何临时存储部分数据？

Question

I have parts of data the coming not by order. Let me explain :

User with id 1 sending data ("phone number") but in the same time user with id 78 sending another data ("home address").

All the incoming data have the same receiver.

When all set of the data per id is set the data would be sent to mysql database and deleted from temporarily storage.

Each user id need to fill 6 different information before the complete data is sent.

So the question is how to to store the temporary data parts using their personal id (without mixing them up) and only then I have all the parts I will proceed to the next task ?

Should I use arrayList or something different ?

Edit (answering the duplicate suggestion ): my question is a bit different and the answer there not helping me at all !

Answer 1

you can create object 'Data'

class Data {

    private param1,param2,...,param6;


    public Data(){

    }

    /* PARAMS getters */

    /* PARAMS setters */


    public void insert(){
        //insert to database
    }

}

then use HashMap to update the data :

HashMap<Integer, Data> map = new HashMap<Integer, Data>();

and every time you recive some data and id check the type of the data (which param in the Data object) and update it in the map

Data toUpdate = map.get(id);
toUpdate.setParam...

you can make a listener for each time you set some param to indicate that all the data had been set and ready to be inserted

Answer 2

You could, as you say, store it in some structure in memory. I'd pick some method which isn't tied to an Activity, otherwise you need to track the activity lifecycle. Eg you could have a HashMap to map entity name (ie name, address, etc) to its value (or in case you're storing multiple instances of a single entity, go with some kind of a Data class suggested in the other answer). Personally, I find this method most cumbersome/cluttered, because you need to either divorce your storage from your Activity or handle all activity changes in order not to lose state.

Second approach is using SharedPreferences. Make a separate prefs file for that purpose, obtain them ( getSharedPreferences("partial_data_dl", MODE_PRIVATE) ), store each column as one field in the prefs, and clear them before commiting them to the database. It's easier than storing everything in memory and shouldn't be noticeably slower.

Third is the obvious one: why don't you update the database record as the data comes in? First time insert a new record with only one column populated, every other issue an update query to add value for the new column. This is admittedly the heaviest solution and might not work for your use case, but I'd give it a try and test it out. As the saying goes, premature optimization is the root of all evil.