Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

PRQL: Pipelined Relational Query Language

github.com

Post reply on HN
Print view View original

211–214 of 214 posts

Re: PRQL: Pipelined Relational Query Language

#211

Post by johnthescott »

inopinatus wrote: post #178
jiggawatts wrote: post #172

There are about 17 of these "SQL but better if we start with the FROM statement at the top" languages. The language-integrated query (Linq) feature of .NET works the same away: var studentsGroupByStandard = from s in studentList group s by s.StandardID into sg orderby sg.Key select new { sg.Key, sg }; Another example is Microsoft's Kusto Query Language, which they use in Azure Log Analytics: StormEvents | where State…

So did QUEL, as implemented in Ingres circa 1976. Perhaps this is what happens when you go back to basics and take a good long look at Codd's relational calculus (or, equivalently, relational algebra), to which SQL bears only a passing resemblance.

back in my day, i thought a WHERE clause on the "RANGE OF alias IS RELATION" would be useful. nowadays, a session level VIEW is the way to go.

Re: PRQL: Pipelined Relational Query Language

#212

Post by snthpy »

tabtab wrote: post #208

Except for niches, one is not going to unseat SQL as the primary de-facto standard relational query language unless the replacement is significantly better, not just slightly. Unless a fad storm kicks in, being slightly better is rarely enough to unseat an established standard. As far as SQL competitor candidates, my personal favorite is SMEQL. It's more composable and uses tables for many operations instead of dedic…

Thanks. I hadn't heard of SMEQL before. To me it structurally actually looks very similar to PRQL.

PRQL just always has `srt` as an implicit parameter (now referenceable as `this` since 0.9) and a more linear flow whereas with SMEQL I guess you could have more complex chains. Actually you can do those with PRQL as well, you just need to break them out into separate `let` expressions. On the upside, PRQL can do away with a lot of the parentheses and some punctuation which makes it a bit nicer IMHO but YMMV.

I agree with you that PRQL would need to be *significantly* better to justify the time investment in learning it. For some of that is already the case but we'll continue working on it to try and hit that sweetspot for even more people.

All the best!

(Disclaimer: I'm a PRQL contributor.)

Re: PRQL: Pipelined Relational Query Language

#213

Post by snthpy »

anon84873628 wrote: post #153

Has anyone compared PRQL and Malloy head to head?

I'm not aware of a head-to-head comparison. My impression is that they have slightly different aims.

PRQL intends to express analytical queries as a series of transformations which form a data pipeline.

My sense is that Malloy aims more to provide a kind of semantic layer as well as dealing with nested data and reports.

(Disclaimer: I'm a PRQL contributor.)

Re: PRQL: Pipelined Relational Query Language

#214

Post by tabtab »

snthpy wrote: post #212
tabtab wrote: post #208

Except for niches, one is not going to unseat SQL as the primary de-facto standard relational query language unless the replacement is significantly better, not just slightly. Unless a fad storm kicks in, being slightly better is rarely enough to unseat an established standard. As far as SQL competitor candidates, my personal favorite is SMEQL. It's more composable and uses tables for many operations instead of dedic…

Thanks. I hadn't heard of SMEQL before. To me it structurally actually looks very similar to PRQL. PRQL just always has `srt` as an implicit parameter (now referenceable as `this` since 0.9) and a more linear flow whereas with SMEQL I guess you could have more complex chains. Actually you can do those with PRQL as well, you just need to break them out into separate `let` expressions. On the upside, PRQL can do away w…

I frankly don't see many similarities to PRQL.
Post reply on HN