getquill/quill

Compile-time Language Integrated Query for Scala

Michael Genereux

@mgenereu

Since my whole stack is based on Li's code, I'm waiting too. ;-)

Matthew de Detrich

@mdedetrich

@deusaquilus @juliano new PR here, I have decided to make a separate PR to remove the redundant null checks that I talked about earlier.

I will then rebase my datetime DSL PR once this is merged into master

Matthew de Detrich

@mdedetrich

Also I am getting a lot of CI errors, is there something wrong with travis currently?

Matthew de Detrich

@mdedetrich

Okay, getquill/quill#1648 is ready to merge!

Juliano Alves

@juliano

merged! Thanks @mdedetrich :)

Li Haoyi

@lihaoyi

is there an easy way to ask for an untyped "resultset" or "row" of stuff in quill?

e.g. I want to do ctx.run(infix"select * from information_schema.tables".as[???]) and have it show me what tables exist, but I don't care what format it is in, as long as I can read it

Li Haoyi

@lihaoyi

I’m tryying to do the basic “list tables” “list columns” stuff that you normally do in psql, but do it in quill instead

Kostya Golikov

@lazyval

Hi! Is there a way to specify a fetch size without using quill-jdbc-monix?

Matthew de Detrich

@mdedetrich

What do you mean by fetch size?

@lihaoyi Can't you just do exists?

i.e.

scala ctx.run(query[InfromationSchema.Tables].exists)`

Kostya Golikov

@lazyval

I mean the hint that you usually specify in jdbc driver / statement level to say how much rows to fetch in one round trip.
https://stackoverflow.com/a/15559710

there was a discussion in gitter previously, as some people wanted to set this hint to 1, to fetch one row at a time, but my use-case is the opposite: I know, in advance, that a specific query gonna return a lot of rows and when I profile it I see there is a lot of roundtrips done to the db, so I would love to bump it somehow from the default value of 10.

todd-son

@todd-son

I want to use raw sql in quill.
Invalid sql was executed.

SELECT x._1, x._2 FROM (SELECT NAME, AGE FROM EAMON_BATCH_PERSON WHERE NAME = 'todd')

My code is

val selectSql = quote {
      (name: String) => infix"""SELECT NAME, AGE FROM EAMON_BATCH_PERSON WHERE NAME = $name""".as[Query[(String, Int)]]
    }

    val result: List[(String, Index)] = run {
      selectSql("todd")
    }

Alias for columns was not generated.
Quill jdbc version is 3.4.9

Li Haoyi

@lihaoyi-databricks

found a new crasher if anyone wants to look :) getquill/quill#1655

Li Haoyi

@lihaoyi-databricks

or if anyone has any hints for workarounds

seems like a pretty straightforward query, so not sure what I can do to shuffle it around and make the macros happy

Li Haoyi

@lihaoyi-databricks

after I define a implicit val encodeInstant: Encoder[java.time.Instant] to allow serializing and deserializing instants to my database, is there a way to perform timestamp arithmetic in the database using Quill?

Matthew de Detrich

@mdedetrich

is there a way to perform timestamp arithmetic in the database using Quill?

I am in the process of writing a DSL for this

Regarding #1655 , I think the issue is with lift(now) - r.timestamp > s._2

Typically I had to create custom infix's to support this, i.e. see https://github.com/mdedetrich/quill-interval-bug/tree/fix/src/main/scala

Li Haoyi

@lihaoyi

thanks, i’ll use an infix for now

Alexander Ioffe

@deusaquilus

@all 3.4.10 is out with support for Scala 2.13. Much thanks to @jilen for making this happen!

Li Haoyi

@lihaoyi

:+1:

Oleg Pyzhcov

@oleg-py

Hey guys, anybody saw something like this:

[error] scala.reflect.macros.TypecheckException: package com.testproject.data is not a value
[error]         at scala.reflect.macros.contexts.Typers.$anonfun$typecheck$3(Typers.scala:44)
[error]         at scala.reflect.macros.contexts.Typers.$anonfun$typecheck$2(Typers.scala:38)
[error]         at scala.reflect.macros.contexts.Typers.doTypecheck$1(Typers.scala:37)
[error]         at scala.reflect.macros.contexts.Typers.$anonfun$typecheck$7(Typers.scala:50)
[error]         at scala.reflect.internal.Trees.wrappingIntoTerm(Trees.scala:1756)
[error]         at scala.reflect.internal.Trees.wrappingIntoTerm$(Trees.scala:1753)
[error]         at scala.reflect.internal.SymbolTable.wrappingIntoTerm(SymbolTable.scala:28)
[error]         at scala.reflect.macros.contexts.Typers.typecheck(Typers.scala:50)
[error]         at scala.reflect.macros.contexts.Typers.typecheck$(Typers.scala:32)
[error]         at scala.reflect.macros.contexts.Context.typecheck(Context.scala:18)
[error]         at scala.reflect.macros.contexts.Context.typecheck(Context.scala:18)
[error]         at io.getquill.quotation.Parsing$$anonfun$propertyParser$1.applyOrElse(Parsing.scala:488)
[error]         at io.getquill.quotation.Parsing$$anonfun$propertyParser$1.applyOrElse(Parsing.scala:474)

In my case com.testproject.data is a package that has package object with some type aliases, and it comes from a different subproject entirely

Oleg Pyzhcov

@oleg-py

seems to be caused by usage of a constant enum value (from enumeratum) without lift

Filipe Moura

@filipe.moura_gitlab

How can I sort by a column from a option left join?

Matthew de Detrich

@mdedetrich

@oleg-py Indeed, we don't support lifting constant values from unsupported types

In order to do this you would have to make a system for macros that would, at compile time, evaluated the enum straight into its relevant sql type which would require changes to both io.quill.Encoder and io.quill.Decoder

Oleg Pyzhcov

@oleg-py

Sounds ok to me, but I'd love a better compile error

Matthew de Detrich

@mdedetrich

Sure, can you create an issue for it so we can track it?

Filipe Moura

@filipe.moura_gitlab

Hi all, How can I sort by a column from a option left join? Any help?

Leo

@leobenkel

Hello, are with clause supported ?

dvir

@dvirf1

:point_up: October 8, 2019 7:07 AM
Thanks @jilen :)

Hubert Słojewski

@Hoobie

Hi, I have a question about Cassandra support. How to perform operations on the list type, using existing elements, is it possible to filter them?
sth like this doesn't work

update(
  report => {
    report.workChunks -> report.workChunks
      .filter(_.uuid != lift(uuid))
    },
)

Li Haoyi

@lihaoyi-databricks

Hi! I'm getting an error with the following update:

    svc.ctx.run{
      query[db.Run]
        .filter(r => liftQuery(timedOutWorkers.map(_.id)).contains(r.workerId))
        .foreach{r =>
          query[db.RunEvent].insert(
            _.runId -> r.id,
            _.timestamp -> lift(Instant.now()),
            _.status -> lift(db.RunStatus.aborted)
          )
        }
    }

It's saying "Batch actions must be static quotations". Any idea if there's a way to workaround this?

basically I want to do a filter, and then for the results of the filter insert things into a separate table

Oleg Pyzhcov

@oleg-py

You can probably try something dumb:

            _.timestamp -> infix"NOW()".as[Instant],
            _.status -> infix"'aborted'".as[RunStatus]

or maybe you can do query[...].filter[...].map(e => (e, lift(Instant.now()), lift(db.RunStatus.aborted))).foreach { ... }

depends on semantics you want from that Instant.now, I guess

Li Haoyi

@lihaoyi-databricks

interesting, yeah that should work

Li Haoyi

@lihaoyi-databricks

@deusaquilus do you think it would be possible to separate Quill's database-connection-and-transaction context and its compile-time-code-generation context into two separate objects?

I ask because the fact that the compile-time stuff is on the same ctx that I have to dependency-inject all over the place causes tons of problems e.g. due to path-dependent types not lining up

but in reality all I need to do is pass around an untyped database connection; all the compile-time/type-related stuff never changes across my codebase and really could be a global object

I know Quill can be configured to just generate SQL strings, so in theory I should be able to configure it to do the SQL-string part using a global object and only pass around the database connection in ctx

but not sure how hard that would be in practice

Where communities thrive

People

Repo info

Activity