nooberfsh / prusto

A presto/trino client library written in rust.

Prusto

A presto/trino client library written in rust.

Installation

# Cargo.toml
[dependencies]
prusto = "0.5"

In order to use this crate as presto client, enable "presto" feature.

# Cargo.toml
[dependencies]
prusto = { version = "0.5", features = ["presto"] }

Example

use prusto::{ClientBuilder, Presto};

#[derive(Presto, Debug)]
struct Foo {
    a: i64,
    b: f64,
    c: String,
}

#[tokio::main]
async fn main() {
    let cli = ClientBuilder::new("user", "localhost")
        .port(8090)
        .catalog("catalog")
        .build()
        .unwrap();

    let sql = "select 1 as a, cast(1.1 as double) as b, 'bar' as c ";

    let data = cli.get_all::<Foo>(sql.into()).await.unwrap().into_vec();

    for r in data {
        println!("{:?}", r)
    }
}

License

MIT

prusto's Issues

Make Client::get public

This method shouldn't be private, it is useful to be able to access full result for custom usecases. Happy to make PR, currently have to fork.

https://github.com/nooberfsh/prusto/blob/master/src/client.rs#L426

The package depends on prusto, with features: presto but prusto does not have these features

Hello,

I am trying to run the example code in the main markdown.

I have installed rustup 1.25.1, rustc 1.68.0-nightly and my toml is

prusto = { version = "0.4.0", features = ["presto"] }

When I cargo run it, I get

failed to select a version for prusto
versions that meet the requirements ^0.4.0 (locked to 0.4.0) are: 0.4.0
the package depends on prusto, with features: presto but prusto does not have these features.

Am I getting something wrong?

Split trino model in a single crate

Do you think it could be a good idea to split your trino model into a single crate?
This will allow people to create clients and implementations using the model

review current type system

How do you deal with gzip?

I found your crate and I couldn't figure out how you are decompressing the gzip body from the response. Is that done automatically by reqwest?

Feat: Get column names without specifying struct from select * statement

When calling select * from table; we are getting only the data but no column names. The suggestion is to get also the column names with their data.

should return error if data set is empty ?

support &str

add execute to client

SHOW CREATE MATERIALIZED VIEW gives empty data

hello,
I need to execute this query
SHOW CREATE MATERIALIZED VIEW view_name

I'm using method get_all, I can see from trino's console that the query is executed and finish properly
but the client gives me empty_data error

I tried both using a String and a struct containing a string as model to map the result but no luck.
how can I achieve this?

thank you

review current data types

update tokio to 0.3

How to add custom header to queries?

Hi

I need to add a custom HTTP header to my queries, is this possible with the current code base?

数据表中如果有 array 字段会反序列化失败， 但是表现的结果是返回0条数据

add a custom stream type

• add a close method

add doc

Allow self-signed certificates.

Our testing/dev environment for trino is using a "self-signed" certificate. Would you consider allowing unverified certificates in reqwest client?

I have created a fork that works for our setup: https://github.com/nooberfsh/prusto/compare/master...siudej:prusto:master?expand=1

add integration tests

Selecting an optional type results in None?

I have a field that can be Null and I've wrapped the type in an Option, something like:

#[derive(prusto::Presto)]
struct MyRow {
  val: Option<String>,
}

When I select it, I get back no rows! If I remove this field and try again I get the rows.

support tuple

remove openssl dependency

Support more than 16 fields for Structs

The following struct compiles correctly:

#[derive(Presto)]
struct Foo {
  a1: i16,
  a2: i16,
  a3: i16,
  a4: i16,
  a5: i16,
  a6: i16,
  a7: i16,
  a8: i16,
  a9: i16,
  a10: i16,
  a11: i16,
  a12: i16,
  a13: i16,
  a14: i16,
  a15: i16,
  a16: i16,
}

If we add a17: i16 as an extra struct member we get the following error:

error[E0277]: the trait bound `(&'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16, &'_a i16): Serialize` is not satisfied
  --> src/main.rs:13:10
   |
13 | #[derive(Presto)]
   |          ^^^^^^ the trait `Serialize` is not implemented for `(&i16, &i16, &i16, &i16, &i16, &i16, &i16, &i16, &i16, &i16, &i16, &i16, &i16, &i16, &i16, &i16, &i16)`
   |
   = help: the following other types implement trait `Serialize`:
             ()
             (T0,)
             (T0, T1)
             (T0, T1, T2)
             (T0, T1, T2, T3)
             (T0, T1, T2, T3, T4)
             (T0, T1, T2, T3, T4, T5)
             (T0, T1, T2, T3, T4, T5, T6)
           and 9 others
note: required by a bound in `prusto::Presto::ValueType`
  --> /Users/bogdanstate/.cargo/registry/src/index.crates.io-6f17d22bba15001f/prusto-0.5.1/src/types/mod.rs:69:25
   |
69 |     type ValueType<'a>: Serialize
   |                         ^^^^^^^^^ required by this bound in `Presto::ValueType`
   = note: this error originates in the derive macro `Presto` (in Nightly builds, run with -Z macro-backtrace for more info)

I assume this is due to a the way the derive(Presto) macro is implemented, i.e. by using a tuple as the underlying type.

add more method for DataSet, e.g. len()

"simple" example appears to not work: always "empty data"

Create a Trino database with some data in it, and try to select it out with get_all. It doesn't work?

#[tokio::main]
async fn main() {
    let user = "admin".to_string();
    let host = "localhost".to_string();
    let port = 8080;
    let catalog = "iceberg".to_string();
    let sql = "SELECT event_id, timestamp, message FROM simba.events".to_string();

    let cli = ClientBuilder::new(user, host)
        .port(port)
        .catalog(catalog)
        .build()
        .unwrap();

    let data = cli.get_all::<Row>(sql).await.unwrap().into_vec();

    for r in data {
        println!("{:?}", r)
    }
}

Did some digging, and I can see the data coming back in the "data" field, but that doesn't make it into the QueryResult deserialization:

{"id":"20240424_091127_00190_vnqi9","infoUri":"http://localhost:8080/ui/query.html?20240424_091127_00190_vnqi9","partialCancelUri":"http://localhost:8080/v1/statement/executing/partialCancel/20240424_091127_00190_vnqi9/0/yb90d84f2d580220289082106b260e8edfec120c2/1","nextUri":"http://localhost:8080/v1/statement/executing/20240424_091127_00190_vnqi9/yb90d84f2d580220289082106b260e8edfec120c2/1","columns":[{"name":"event_id","type":"varchar","typeSignature":{"rawType":"varchar","arguments":[{"kind":"LONG","value":2147483647}]}},{"name":"timestamp","type":"timestamp(6) with time zone","typeSignature":{"rawType":"timestamp with time zone","arguments":[{"kind":"LONG","value":6}]}},{"name":"message","type":"varchar","typeSignature":{"rawType":"varchar","arguments":[{"kind":"LONG","value":2147483647}]}}],"data":[["w-1","2024-04-24 22:05:45.728000 UTC","what the fuck"],["w-2","2024-04-24 22:05:45.728000 UTC","are we doing"],["w-3","2024-04-23 22:05:45.728000 UTC","what the fuck here, also"]],"stats":{"state":"RUNNING","queued":false,"scheduled":true,"progressPercentage":100.0,"runningPercentage":0.0,"nodes":1,"totalSplits":1,"queuedSplits":0,"runningSplits":0,"completedSplits":1,"cpuTimeMillis":9,"wallTimeMillis":10,"queuedTimeMillis":1,"elapsedTimeMillis":12441,"processedRows":3,"processedBytes":258263,"physicalInputBytes":258263,"physicalWrittenBytes":0,"peakMemoryBytes":590,"spilledBytes":0,"rootStage":{"stageId":"0","state":"RUNNING","done":false,"nodes":1,"totalSplits":1,"queuedSplits":0,"runningSplits":0,"completedSplits":1,"cpuTimeMillis":9,"wallTimeMillis":10,"processedRows":3,"processedBytes":258263,"physicalInputBytes":258263,"failedTasks":0,"coordinatorOnly":false,"subStages":[]}},"warnings":[]}

QueryResult {
    id: "20240424_091127_00190_vnqi9",
    info_uri: "http://localhost:8080/ui/query.html?20240424_091127_00190_vnqi9",
    partial_cancel_uri: Some(
        "http://localhost:8080/v1/statement/executing/partialCancel/20240424_091127_00190_vnqi9/0/yb90d84f2d580220289082106b260e8edfec120c2/1",
    ),
    next_uri: Some(
        "http://localhost:8080/v1/statement/executing/20240424_091127_00190_vnqi9/yb90d84f2d580220289082106b260e8edfec120c2/1",
    ),
    data_set: None,
    error: None,
    stats: Stat {
        state: "RUNNING",
        queued: false,
        scheduled: true,
        nodes: 1,
        total_splits: 1,
        queued_splits: 0,
        running_splits: 0,
        completed_splits: 1,
        cpu_time_millis: 9,
        wall_time_millis: 10,
        queued_time_millis: 1,
        elapsed_time_millis: 12441,
        processed_rows: 3,
        processed_bytes: 258263,
        peak_memory_bytes: 590,
        spilled_bytes: 0,
        root_stage: None,
    },
    warnings: [],
    update_type: None,
    update_count: None,
}

Would it be possible to release 0.5.1?

Hello! I'm experimenting with Trino and Rust, and your crate is very helpful!

I noticed that you fixed a bug in ClientBuilder::schema in 3d0f011. Would it be possible to get an 0.5.1 release of prusto with that fix at some point? (I can use a git version in Cargo.toml, but that would prevent me from ever making an open source release using crates.io.)

Thank you very much for maintaining a Trino/Presto client!

set session from response