FWD » Core Development » Database

01/29/2015

02:46 PM Bug #2488: replace runtime compilation of dynamic queries

This update fixes the copy/paste bug revealed by issues #2500 and #2501. Ovidiu Maxiniuc

01/28/2015

04:21 PM Bug #2488 (WIP): replace runtime compilation of dynamic queries

Added 'dumb' cache implementation and some additional fixes. Ovidiu Maxiniuc

01/26/2015

03:14 AM Bug #2488: replace runtime compilation of dynamic queries

Eric,
If you did not started the review, please find attached an update that adds the missing javadocs to @RuntimeJa... Ovidiu Maxiniuc

01/23/2015

09:29 AM Bug #2488: replace runtime compilation of dynamic queries

Greg, you are right.
We do parse the initial query using the @progress.g@ (@ProgressParser@ from the code in note ... Ovidiu Maxiniuc

08:58 AM Bug #2488: replace runtime compilation of dynamic queries

When I refer to the Progress AST, I mean the original AST that is first created for the provided expression in a @QUE... Greg Shah

08:55 AM Bug #2488: replace runtime compilation of dynamic queries

A few thoughts on the caching:
I agree with Greg that the Progress AST will be much more reliable. However, it is ... Eric Faulhaber

08:53 AM Bug #2488: replace runtime compilation of dynamic queries

Not sure I understand. Of course, the trees will be the except for the string value at some point. But later, during ... Ovidiu Maxiniuc

08:26 AM Bug #2488: replace runtime compilation of dynamic queries

> and realize that they are the same ?
Don't forget that in the Progress AST the tree structure is in fact:
* ... Greg Shah

05:01 AM Bug #2488: replace runtime compilation of dynamic queries

I analyzed the logs with the queries from #1868. Here are some results/thoughts:
* there are ~5500 @FIND FIRST@ of w... Ovidiu Maxiniuc

01/22/2015

04:56 PM Bug #2488: replace runtime compilation of dynamic queries

Uploaded update for review, together a file with my important testcases.
* I tried to detect as much as possible i... Ovidiu Maxiniuc

06:02 AM Bug #2488: replace runtime compilation of dynamic queries

After I got satisfactory results with simpler queries, I started to increase the complexity of the dynamic queries in... Ovidiu Maxiniuc

03:38 PM Bug #2496: Static conversion of legacy field names does not preserve original case

Legacy field names annotations for static conversion are now in low case. TableMapper seems works properly, also ther... Vadim Nebogatov

03:07 PM Bug #2496 (New): Static conversion of legacy field names does not preserve original case

Vadim Nebogatov

01/21/2015

10:28 AM Bug #2488: replace runtime compilation of dynamic queries

Eric Faulhaber wrote:
> Don't we convert these kinds of expressions into HQL today, using PL/Java UDFs to represent ... Ovidiu Maxiniuc

10:27 AM Bug #2488: replace runtime compilation of dynamic queries

Ovidiu Maxiniuc wrote:
> On the other hand, if we do to many semantic annotations, the tree will grow - and the mana... Eric Faulhaber

10:19 AM Bug #2488: replace runtime compilation of dynamic queries

Ovidiu Maxiniuc wrote:
> The static call is emitted, relying on java compiler to resolve the actual method to be inv... Eric Faulhaber

02:58 AM Bug #2488: replace runtime compilation of dynamic queries

This is true, we don't do the constant expression evaluation at this moment neither in static conversion but it would... Ovidiu Maxiniuc

01/20/2015

04:36 PM Bug #2488: replace runtime compilation of dynamic queries

Certainly static expressions will be more efficient to convert, execute, and store at conversion. I'm pretty sure we ... Eric Faulhaber

03:55 PM Bug #2488: replace runtime compilation of dynamic queries

> Don't abandon it just yet...it will not need to do as much as it does now, but this is a good candidate to morph in... Ovidiu Maxiniuc

08:08 PM Bug #2488: replace runtime compilation of dynamic queries

Ovidiu Maxiniuc wrote:
> Attached you can find a first version of 'runtime/interpret_find_query.xml'. It only handle... Eric Faulhaber

01/19/2015

02:36 PM Bug #2488: replace runtime compilation of dynamic queries

Attached you can find a first version of 'runtime/interpret_find_query.xml'. It only handles the simple @find-query@ ... Ovidiu Maxiniuc

08:07 PM Bug #2488: replace runtime compilation of dynamic queries

Ovidiu Maxiniuc wrote:
> If you have the chance to re-run it, please add the string @predicate@ before and, to simpl... Eric Faulhaber

01/17/2015

05:27 PM Bug #2488: replace runtime compilation of dynamic queries

Eric Faulhaber wrote (from Constantin's email):
> The problem here is that at conversion-time the SchemaDictionary (... Eric Faulhaber

01/16/2015

05:46 PM Bug #2488: replace runtime compilation of dynamic queries

BTW, we have implemented a lot of constructor and method matching support already for TRPL itself. Perhaps we can lev... Eric Faulhaber

05:37 PM Bug #2488: replace runtime compilation of dynamic queries

I've taken a look at the code so far and it is a great start! Yes, I agree that the final tree is simpler to deal wit... Eric Faulhaber

03:55 PM Bug #2488: replace runtime compilation of dynamic queries

Eric Faulhaber wrote:
> Sure, I can do that, but are you certain you want to use the current, final version of the J... Ovidiu Maxiniuc

03:41 PM Bug #2488: replace runtime compilation of dynamic queries

Ovidiu Maxiniuc wrote:
> I did some researches on windev01. Progress is not very permissive. The static statements c... Eric Faulhaber

03:30 PM Bug #2488: replace runtime compilation of dynamic queries

Related to the @createAlias@ topic, here's an email exchange I had with Constantin recently:
> Eric,
>
> The prob... Eric Faulhaber

03:28 PM Bug #2488: replace runtime compilation of dynamic queries

> The text of the error message suggests you are passing the same parameter (@dictdb@) for both the @alias@ and @ldbN... Ovidiu Maxiniuc

02:38 PM Bug #2488: replace runtime compilation of dynamic queries

While I am excited that you've begun implementing a solution, I'm curious whether you've found the answers to the que... Eric Faulhaber

02:31 PM Bug #2488: replace runtime compilation of dynamic queries

Ovidiu Maxiniuc wrote:
> I started implementing a small 'interpreting' engine, targeting only the required node/item... Eric Faulhaber

02:25 PM Bug #2488: replace runtime compilation of dynamic queries

I started implementing a small 'interpreting' engine, targeting only the required node/items. In fact at this time, t... Ovidiu Maxiniuc

01/15/2015

03:12 PM Bug #2488: replace runtime compilation of dynamic queries

Ovidiu Maxiniuc wrote:
> Yes, @query-prepare()@ is the big issue. I believe there are no constrains here and any num... Eric Faulhaber

02:58 PM Bug #2488: replace runtime compilation of dynamic queries

Yes, @query-prepare()@ is the big issue. I believe there are no constrains here and any number of buffers can be adde... Ovidiu Maxiniuc

02:53 PM Bug #2488: replace runtime compilation of dynamic queries

OK, so no calls to UDFs, either? That's another way to trigger client-side where clause processing.
If not, it wou... Eric Faulhaber

02:42 PM Bug #2488: replace runtime compilation of dynamic queries

Looking at @find-first()@ method in manual, the predicate expression is very simple, just ... Ovidiu Maxiniuc

02:34 PM Bug #2488: replace runtime compilation of dynamic queries

I think the following would give us a reasonably complex @WhereExpression@:... Eric Faulhaber

01:05 PM Bug #2488: replace runtime compilation of dynamic queries

> I'm not sure how well an interpreted WhereExpression would work. We may still have to fall back to a compiled appro... Ovidiu Maxiniuc

12:47 PM Bug #2488: replace runtime compilation of dynamic queries

Ovidiu Maxiniuc wrote:
> The java/TRLP debate. It's clear that TRPL has the some advantages, but I think that we sho... Eric Faulhaber

12:41 PM Bug #2488: replace runtime compilation of dynamic queries

Ovidiu Maxiniuc wrote:
> The difficulty appears when there are another buffers (and aliases)...
Please explain your... Eric Faulhaber

06:41 AM Bug #2488: replace runtime compilation of dynamic queries

I understand the principle of this approach.
At first sight, it looks rather easy, the where clause and sorting ar... Ovidiu Maxiniuc

02:04 AM Bug #2488: replace runtime compilation of dynamic queries

The basic idea behind this improvement is to continue to use runtime TRPL to convert a query to an in-memory, Java AS... Eric Faulhaber

01/14/2015

01:26 AM Bug #2488: replace runtime compilation of dynamic queries

Dynamically defined queries are parsed and converted at runtime by P2J, then compiled into discrete Java classes usin... Eric Faulhaber

12:48 AM Bug #2488 (Closed): replace runtime compilation of dynamic queries

Eric Faulhaber

01/09/2015

08:02 AM Bug #2482 (WIP): conversion of FIND does not honor NO-LOCK when specified early

Vadim Nebogatov

02:29 AM Bug #2482: conversion of FIND does not honor NO-LOCK when specified early

I found that a FIND statement of the following form is converted without a @LockType.NONE@ lock type:... Eric Faulhaber

02:03 AM Bug #2482 (WIP): conversion of FIND does not honor NO-LOCK when specified early

Eric Faulhaber