I'm trying to ingest a sequence of 6293 strings, each 8 characters long, and create a map in which each key is an integer from 1 to 63, and each value is a subsequence of the 6293 strings, from ($key+1) * 100) to that starting value + 99.
I wrote a function to do this recursively (so I thought), but it runs and runs (I've given it over 5 minutes), and so I can see I borked something. I wrote a function to output the starting and ending sequence indices for each execution of the function, but it (as the only change to the first function) quickly causes a stack overflow:
*Error:* Stack Overflow: Try tail recursion?
Since the recursive call was the last expression in the function, I thought I was doing tail recursion.
I'm running 11.6.0. The code with explanatory comments follows. I'd be grateful for any corrections and insights.
declare function local:write_results($elems as element()*, $count as xs:integer) { let $dir_path := '/home/cfbearden/projects/mesh_for_pure/20241220/data/' let $count_as_str := xs:string($count) let $content_out := <PubmedArticleSet>{$elems}</PubmedArticleSet> let $path := $dir_path || 'pmed2022_' || $count_as_str || '.xml' let $params := { 'method' : 'xml', 'indent' : 'yes' } return file:write($path, $elems, $params) };
declare function local:by_hundreds($pmids as xs:string*, $pmid_map as map(*)) as map(*) { let $key_count := count(map:keys($pmid_map)) return (: base case: each value except the last should have 100 items in it; the last should have <= 100; in this case we return the map goal is to build a map with 63 keys, with each value a sequence of 100 or 93 (in the last case) strings :) if ((($key_count + 1) * 100) >= count($pmids)) then $pmid_map else (: starting index for subsequence() :) let $start := 1 + ($key_count * 100) (: ending index for subsequence() :) let $end := $start + 99 (: value of the map key for this 100 items :) let $key := $key_count + 1 (: attempt to log all recurrent executions to files let $foo := local:write_results(<res><start>{$start}</start><end>{$end}</end></res>, $key_count) :) (: create a map with $key as key and the subsequence from n to n + 99 of $pmids and merge it with input accumulator map to make a new map :) let $pmid_map_new := map:merge($pmid_map, map { $key : subsequence($pmids, $start, $end) }) (: recur; the number of keys in $pmid_map_new will enable the next execution to determine which items from $pmids to use :) return local:by_hundreds($pmids, $pmid_map_new) };
(: The data format is simple: <DATA_RECORDS> <DATA_RECORD> <pub_uuid>0008980a-f1f3-40bf-8c13-cb0b79dfa81e</pub_uuid> <pmid>35732127</pmid> </DATA_RECORD> ... </DATA_RECORDS> :) let $pure_2022_db := collection('pure_2022')
(: yields a sequence of 6293 strings, all 8 characters long :) let $pmids := ( for $drec in $pure_2022_db/DATA_RECORDS/DATA_RECORD let $pmid := $drec/pmid/text() return $pmid )
let $result := local:by_hundreds($pmids, {}) return $result
All the best, Chuck Bearden