@conan Looks really nice. Thanks!

morning

Anyone suggest coding kata sized challenges (1-4 hours long) for those just starting clojure / clojurescript? Links to simple solutions a bonus. Thanks.

I always like to plot a mandlebrot or julia set when I first start in a language - you can do that in under an hour, but you can also spend more than 4 hours on it.

it does require some maths and maybe a headstart with graphics code for beginners

final solution is usually quite simple; one nested for loop over the coordinates (x & y), one iteration loop doing the (in fact quite simple) maths, and one drawing command + language-dependant overhead for actually spitting the fractal out somewhere

Morning ☀️

Morning All…

🙂

@conan - I had no idea that ClubHouse was a Clojure app; how cool 🙂

no i've been using it for a couple of months and only just found out myself!

So, lovely UK Clojure people, I have a question… If you had this function:

(defn init-noaa-canonical-db-pt
  "Command Function to create PostgreSQL databse for NOAA daily observations data - PT"
  []
  (let [st-data (rest (ingest-csv-data "/rawdata/NOAA_DailySummaries_Portugal_1992_to_2016_normalised.csv"))
        st-data-prepared (get-station-data-prepared st-data (:station-keys ingest-cfg))]
    (doseq [row st-data-prepared]
      (sql/populate-noaa-stations-pt! {:? (ordered-values-vector-from-map
                                        (:station-keys ingest-cfg)
                                        row)}))
    (doseq [row st-data]
      (sql/populate-noaa-daily-measurements-pt! {:? (ordered-values-vector-from-map
                                                  (:noaa-measurement-keys ingest-cfg)
                                                  (first (noaa-prepare-measurements-data
                                                          row)))}))))

(I am using Yesql, so there are different functions for each table Insert, i.e. sql/populate-noaa-stations-pt! as above for Portugal, and as an example sql/populate-noaa-stations-ci! for Côte d’Ivoire)

i.e. Can one use a macro to re-write the symbol of a function call or several function calls?

Does Yesql let you enumerate the functions it generates in some way?

(Clearly the string param could become a config value)

@conan - Not that I am aware of, and I have spent some time spelunking in the docs, but I may__ be wrong…

(i.e. probably am)

you could probably just read the files from your sql directory actully, the naming is based on those, right? then you could generate the symbols

probably

(it's been a while since i used yesql)

In the “here and now” I am prepared to create 5 other, practically identical functions, but overall that much repetition feels wrong.

@conan - I am using the “one file” approach, with annotations / comments that name__ the queries / functions

could you put all the country codes in a db table and iterate over them with the same sql function?

or are they all different

i'm trying to think of anything that is not a macro!

Right, so because Yesql ONLY allows for parameterised replacement of values, i.e. you can’t parameterise a tablename, that would not help - using Yesql I HAVE TO have a query per table.

So looping does not really help, unless I store the query / function names.

I am starting to be of the opinion that I need to migrate to HugSQL as that allows for parameter replacement of SQL Identifiers, like table names.

I don’t really want to write a macro - I mean my question boils down to “can one write a macro that will, based on input params, rewrite the invoked functions?”

(I don’t even know if that’s possible)

i was thinking of doing the parameterisation in the SQL query inside a transaction, by loading the data from a table; I might be off here though. That said, you don't have to use Yesql for everything, in this case you could swap it out

@conan - Indeed, and I am starting to think that HugSQL is a better fit for my use cases

it looks like yesql has a raw-sql function

It does, but that rather defeats the object of using the library imho.

(the object being to segregate Clojure and the SQL)

it does feel like the thin end of the wedge

Yeah

HugSQL would allow me to have a generalised query that can do these Inserts on any of the tables that conform to the structure of the data. All my munging functions are already generalised, and are being re-used, so as long as the CSV data is in the right format / structure I already have re-usable functions that can deal with all the CSVs / data sources.

It’s the problem of having a discrete function to INSERT into each of 6 tables that leads to the repetition.

If I swap Yesql out and put HugSQL in I have a (small) hill to climb in terms of adoption and so even temporarily having a macro seems better as I have a deadline to hit…

Workflow help: I have some code which I'm writing hand-in-hand with some extensions to a library. How best to update the library and have it reflected in the REPL (my code) that uses it?

i.e, can I avoid the lein install song'n'dance.

@mattford - I am sad to say that my experience doesn’t go that deep, but that I’ve got a feeling that the lein install dance may be required…

lein checkouts are probably the best way to work across the two in sync

other two options are: decide what you need from the lib and build in isolation from the actual code throw in an extra (ns) form or similar on your side and write the lib code inline until it’s stablised

@conan - In case you are interested, @danieleneal has offered a really interesting take on my question above ^^ which I am certain will work and is very elegant. Ping me if you’d like to see it and I’ll send you the “solution” 🙂

oh yes, i am interested

@mattford use leiningen checkouts

you can just reload the files in the repl too though

Goes to read https://github.com/technomancy/leiningen/blob/master/doc/TUTORIAL.md#checkout-dependencies

@conan - Will get it working and then send you the overview in a chat… 🙂

great, thanks!

hah! that worked a treat. Awesome thanks.

@danieleneal - Just wanted to say “Thanks!” one more time - my code is working very nicely on live and is currently loading LOADS of data into my production databases 🙂