lifecycleScope.launch will start a coroutine, to make it simple the code inside lifecycleScope.launch will be executed in another thread and it will take some time until inserting data and reading it from database, but println(a) is on the main thread so it will be executed before this line a = it[it.size – 1].name, so your println(a) should be inside lifecycleScope.launch like this:

class MainActivity : AppCompatActivity() {
    private var binding: ActivityMainBinding? = null
    private var a: String = "Hi"

    override fun onCreate(savedInstanceState: Bundle?) {
        binding = ActivityMainBinding.inflate(layoutInflater)
        super.onCreate(savedInstanceState)
        setContentView(binding?.root)

        val somethingDao = SomethingDatabase.getDatabase(this).somethingDao()

        lifecycleScope.launch {
            somethingDao.insert(SomethingModel("Hello"))
            somethingDao.fetchAllSomething().collect {
                a = it[it.size - 1].name
                println(a)
            }
        }
    }
}

Note: take a look on kotlin coroutines to better understand

I am confused about how flow.collect works. Because in the lifecycleScope below I already say that a should be assigned by the value of data in my database. However, the value of a is still the string of “Hi” instead of “Hello”.

class MainActivity : AppCompatActivity() {
    private var binding: ActivityMainBinding? = null
    private var a: String = "Hi"

override fun onCreate(savedInstanceState: Bundle?) {
    binding = ActivityMainBinding.inflate(layoutInflater)
    super.onCreate(savedInstanceState)
    setContentView(binding?.root)

    val somethingDao = SomethingDatabase.getDatabase(this).somethingDao()

    lifecycleScope.launch {
        somethingDao.insert(SomethingModel("Hello"))
        somethingDao.fetchAllSomething().collect {
            a = it[it.size - 1].name
        }
    }

    println(a)

}
}

this is all of the information in my database