From 404a5e82ff008eb1bc8ffc584884e38761e61749 Mon Sep 17 00:00:00 2001
From: Emerson Rocha <rocha@ieee.org>
Date: Thu, 28 Apr 2022 00:57:37 -0300
Subject: [PATCH] 1603_1.py (#29): failed; test renaming before merge dont
 work; maybe do after? refs
 https://github.com/EticaAI/multilingual-lexicography/issues/29#issuecomment-1111707350

---
 officinam/999999999/999999999.lib.sh | 32 +++++++++++++++++++++++-----
 1 file changed, 27 insertions(+), 5 deletions(-)

diff --git a/officinam/999999999/999999999.lib.sh b/officinam/999999999/999999999.lib.sh
index b41e98f..62dd07b 100644
--- a/officinam/999999999/999999999.lib.sh
+++ b/officinam/999999999/999999999.lib.sh
@@ -1225,11 +1225,17 @@ file_merge_numerordinatio_de_wiki_q() {
   # TODO: implement check if necessary to revalidate
   echo "${FUNCNAME[0]} sources changed_recently. Reloading... [$fontem_archivum]"
 
+  # NOTE: explanation on the hotfix +ix_deleteme here:
+  #       https://github.com/EticaAI/multilingual-lexicography/issues/
+  #       29#issuecomment-1111707350
+  #       This may be removed when hxlmerge --replace works with so many
+  #       columns at once.
+
   if [ "$est_non_normale" -eq "1" ]; then
     # We apply 'hxlclean --lower' only on writting systems which this make
     # sence. On this case at least '+is_latn,+is_cyrl'
     hxlrename \
-      --rename='item+conceptum+codicem:#item+rem+i_qcc+is_zxxx+ix_wikiq' \
+      --rename='item+conceptum+codicem:#item+rem+i_qcc+is_zxxx+ix_wikiq+ix_deleteme' \
       "$fontem_q_archivum" |
       hxlclean --lower='#*+is_latn,#*+is_cyrl' \
         >"$fontem_q_archivum_temporarium"
@@ -1237,7 +1243,7 @@ file_merge_numerordinatio_de_wiki_q() {
     # We apply 'hxlclean --lower' only on writting systems which this make
     # sence. On this case at least '+is_latn,+is_cyrl'
     hxlrename \
-      --rename='item+conceptum+codicem:#item+rem+i_qcc+is_zxxx+ix_wikiq' \
+      --rename='item+conceptum+codicem:#item+rem+i_qcc+is_zxxx+ix_wikiq+ix_deleteme' \
       "$fontem_q_archivum" \
       >"$fontem_q_archivum_temporarium"
   fi
@@ -1250,15 +1256,31 @@ file_merge_numerordinatio_de_wiki_q() {
 
   # echo "oi2"
 
-  hxlmerge --keys='#item+rem+i_qcc+is_zxxx+ix_wikiq' \
+  hxlmerge --keys='#item+rem+i_qcc+is_zxxx+ix_wikiq+ix_deleteme' \
     --tags='#item+rem' \
     --merge="$fontem_q_archivum_temporarium" \
     "$fontem_archivum" \
+    | hxlcut --exclude='#item+rem+i_qcc+is_zxxx+ix_wikiq+ix_deleteme' \
     >"$objectivum_archivum_temporarium"
 
+  # BUG: if we use hxlmerge --replace, instead of not be repeated on final
+  #      dataset, we lost all additional column data. This migth be because
+  #      we're far beyond typical number of columns libhxl-python is tested to
+  #      work
+
+  # set -x
+  # hxlmerge --keys='#item+rem+i_qcc+is_zxxx+ix_wikiq' \
+  #   --replace \
+  #   --tags='#item+rem' \
+  #   --merge="$fontem_q_archivum_temporarium" \
+  #   "$fontem_archivum" \
+  #   >"$objectivum_archivum_temporarium"
+  # set +x
+
   sed -i '1d' "${objectivum_archivum_temporarium}"
 
-  rm "$fontem_q_archivum_temporarium"
+  # cp "$objectivum_archivum_temporarium" "$objectivum_archivum_temporarium.tmp"
+  # rm "$fontem_q_archivum_temporarium"
 
   file_update_if_necessary csv "$objectivum_archivum_temporarium" "$objectivum_archivum"
 
@@ -2273,7 +2295,7 @@ actiones_completis_locali() {
 
   if [ -z "$(quaero__ix_n1603ia__victionarium_q "$numerordinatio")" ]; then
     # echo "yay"
-    file_translate_csv_de_numerordinatio_q "$numerordinatio" "0" "0" "1"
+    # file_translate_csv_de_numerordinatio_q "$numerordinatio" "0" "0" "1"
     file_merge_numerordinatio_de_wiki_q "$numerordinatio" "0" "0"
     file_convert_tmx_de_numerordinatio11 "$numerordinatio"
     file_convert_tbx_de_numerordinatio11 "$numerordinatio"