New Exercise: Word count

[marianfoo]

Quite a few errors in solution.
Set status to "WIP" so only maintainers see this on exercism (as documented here)

Source

https://github.com/exercism/javascript/tree/main/exercises/practice/word-count

SAP Unit Tests

ABAP Lint

Few errors

local_testclass_location

Don´t know what that means

./exercises/practice/word-count/zcl_word_count.clas.testclass.abap[1, 7] - Place local testclass "ltcl_word_count" in the testclass include (local_testclass_location) 

check_syntax

May have a problem with the GROUP.

./exercises/practice/word-count/.meta/zcl_word_count.clas.abap[51, 7] - Component "size" not found in structure (check_syntax)

check_syntax

return_structure is a local type.

./exercises/practice/word-count/.meta/zcl_word_count.clas.abap[52, 7] - "return_structure" not found, findTop (check_syntax) 

Run Tests in Docker

Don´t know if my docker run is faulty because there are many errors in the output.

[larshp]

rename file zcl_word_count.clas.testclass.abap to zcl_word_count.clas.testclasses.abap

[larshp]

LOOP AT GROUP BY does not work yet

[marianfoo]

LOOP AT GROUP BY does not work yet

no worries, It's probably better to refactor the solution than wait for the implementation which is surely not that easy, right?

[larshp]

we'll never really know 😅

if the exercise is most easily solved with GROUP BY, then we should wait, but 🤷‍♂️ everything has different solutions

[marianfoo]

ok, i have a look if i find a different solution

but there is also the other error
Is this about VALUE ?

[pokrakam]

If the obvious solution is to use group by then we should hold back, because it means others will try that too.

I had a look at the exercise, and can see a couple of ways. IMHO group by is something more savvy people are likely to try, but that's pure speculation.

[larshp]

@marianfoo can you paste the full code example?

Update: oh, its the original code, guess its a derived error

[marianfoo]

its the code in this PR
https://github.com/exercism/abap/blob/362edd56051138dac9afd7a07e1d92788aebd113/exercises/practice/word-count/.meta/zcl_word_count.clas.abap

[marianfoo]

@marianfoo can you paste the full code example?

Update: oh, its the original code, guess its a derived error

ah ok, would make sense

[marianfoo]

If the obvious solution is to use group by then we should hold back, because it means others will try that too.

I had a look at the exercise, and can see a couple of ways. IMHO group by is something more savvy people are likely to try, but that's pure speculation.

You can try with a different solution :D
In any case, it is not a solution for beginners.

[pokrakam]

You can try with a different solution :D

I didn't mean for us, of course we can find workarounds 🙂

What I meant was caution about putting out something where the probability of people running into issues is high. Maybe put a note in the description to avoid LOOP with GROUP BY if time to resolve is unknown, or hold this in draft if we expect it can be implemented in the next week or so.

[larshp]

perhaps just wait a bit, focus on getting #87 merged, then there are 2 more ready to merge?

might leave some time for GROUP BY to be implemented 😅

[marianfoo]

perhaps just wait a bit, focus on getting #87 merged, then there are 2 more ready to merge?

might leave some time for GROUP BY to be implemented 😅

The two others (#88, #89) are concept exercises which need the concepts created first.
This is a "normal" exercise we can publish if it would work.

[mbtools]

I don't think loop ... group by is common knowledge. Seems complicated.

here's my solution:

  METHOD count_words.
    DATA(clean) = replace( val = to_lower( phrase ) sub = `'` with = `` occ = 0 ).
    clean = replace( val = clean sub = `\n` with = ` ` occ = 0 ).
    clean = replace( val = clean sub = `\t` with = ` ` occ = 0 ).
    clean = replace( val = clean regex = `[^a-z0-9]` with = ` ` occ = 0 ).

    SPLIT condense( clean ) AT ` ` INTO TABLE DATA(words).

    LOOP AT words INTO DATA(word).
      DATA(one_result) = VALUE return_structure( word = word count = 1 ).
      COLLECT one_result INTO result.
    ENDLOOP.
  ENDMETHOD.

this does require a change of the result type i.e. with key word which should be done anyway:

    TYPES:
      return_table TYPE STANDARD TABLE OF return_structure WITH KEY word.

PS: Since it's now an abapGit repo, this needs package.devc.xml and the xml for the class. Best to push from a SAP system using abapGit. Then copy the class files to the .meta folder and add the solution there (using GH or VScode).

PPS: abaplint errors will go away with #157.

[marianfoo]

@mbtools thanks for you solution, i created the abapGit Config
it works in 1909 but unfortunately not here

@larshp can you please help here with the error message

[mbtools]

yeah, I forgot that the collect statement isn't implemented yet
https://github.com/abaplint/transpiler/blob/main/packages/transpiler/src/statements/collect.ts

simple statement in abap but lot's of things happening behind the scenes.

you could replace it with:

DATA(one_result) = VALUE return_structure( word = word count = 1 ).
" COLLECT one_result INTO result.
READ TABLE result ASSIGNING FIELD-SYMBOL(<result>) WITH TABLE KEY word = one_result-word.
IF sy-subrc = 0.
  <result>-count = <result>-count + one_result-count.
ELSE.
  INSERT one_result INTO TABLE result.
ENDIF.

[marianfoo]

@mbtools green 🥳

[mbtools]

just remove tests.toml, results.json

[marianfoo]

just remove tests.toml, results.json

done

[mbtools]

Great 👍

[mbtools]

just a question of how many students will use collect ... @larshp can you deal with the "pressure"?

[marianfoo]

just a question of how many students will use collect ... @larshp can you deal with the "pressure"?

programming in abap for 4 years, never heard of it
So i guess not that many use it?

[marianfoo]

Exercise is only and works with the solution :)