Tobias Brandt

Totally agree about finishing #2723 , but in the meantime how is this different to any of the other type inference that's going on? What are the types of `income`...

As mentioned in #3679, I'm in favour of [d-strings](https://discuss.python.org/t/d-string-vs-str-dedent/35907) rather, probably without the d prefix so that it can apply to s- and f-strings too. To wit: > This is...

Hi, this is a good observation and something I noticed as well when answering your other question. It's not too difficult to add this in the PRQL query but I...

Ah, I misunderstood your requirement. The following query should produce the result you want: ```elm let data = [ {id=1, value=2}, {id=2, value=3}, {id=3, value=4}, {id=6, value=2}, {id=5, value=3}, {id=4,...

> Just a link to a book that compares this: https://eitsupi.github.io/querying-with-prql/timeseries#sec-moving-ave Very cool! I also used an SMA like that in my presentations, as well as an EWMA which you...

Some observations: --- # Attempt 1 The error in the `join` can be resolved by using `that` instead of `my_data` but then you hit a bug with a reference in...

I agree with most of these and just have two questions I think: 1. Like @vanillajonathan asked, what's the difference between `instant` / `moment` and `datetime`? 2. `duration` vs `period`:...

Coming late to this discussion but will give it some thought as well.

A bit laborious but I think it shows that you should have all the tools available to you: ```elm from [{n=1}] loop ( filter n

The naming of the column as `_expr_0` is one error. The other is that the order of the columns in the first part of `table_1` before the `UNION` (as was...