Fix clippy warning about loop that never executes

[alamb]

When the flightsql feature is enabled clippy is failing with this warning

error: this loop never actually loops
   --> src/app/handlers/flightsql.rs:89:45
    |
89  | / ...                   while let Some(maybe_batch) = stream.next().await {
90  | | ...                       match maybe_batch {
91  | | ...                           Ok(batch) => {
92  | | ...                               info!("Batch rows: {}", batch.num_rows());
...   |
103 | | ...                       }
104 | | ...                   }
    | |_______________________^
    |
    = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#never_loop
    = note: `#[deny(clippy::never_loop)]` on by default

This code fixes it

[alamb]

the complaint is that this break means the outer loop stops

I think the breaks here mean effectively this code would only ever retrieve the first batch from the remote flightsql server which doesn't seem like the actual intent

[matthewmturner]

it actually was the intent but only as a short term hack to prevent queries with large number of results from taking a long time to return result. i still need to implement auto pagination logic.

[matthewmturner]

since this is flightsql and no dataframe to work with i had in mind auto appending LIMIT and OFFSET to the queries. do you have any other ideas?

[alamb]

I suggest:

1. initially just signal somehow that the stream is truncated and only a part is shown (we can change the code to make this explicit with a comment to get clippy happy)
2. eventually, I can imagine some much fancier logic that fetches a stream incrementally as the user wants to page through the results (if they do). That would only buffer the rows that haev actually been shown, but would allow the user to scroll through them efficiently

I don't think there is any reason to use LIMIT / OFFSET unless there are a huge number of rows and the offset is large (like OFFSET 1000000 or something). I think handling it on the client side would be a lot easier

[matthewmturner]

I thought i had gotten the LIMIT / OFFSET idea from dbeaver but I just double checked and im not seeing it there so yes i agree it doesnt make sense. My original idea was to just paginate through the stream which I believe is same as your second point so I think were aligned on approach here.

[alamb]

related dicussion #137

[alamb]

I'll try a different approach in this PR

[alamb]

This makes clippy happy by being explicit that only the first batch is needed

[alamb]

🚀