From f80dde248765ac86705f4b0a3e4dac5157b547ae Mon Sep 17 00:00:00 2001
From: vv01f <vv01f@users.noreply.github.com>
Date: Sun, 17 Nov 2019 00:23:10 +0100
Subject: [PATCH] convert to target csv format

---
 .gitignore |  1 +
 dhl.sh     | 33 ++++++++++++++++++++++++++++++---
 2 files changed, 31 insertions(+), 3 deletions(-)

diff --git a/.gitignore b/.gitignore
index d6b3082..ccb40ff 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,2 +1,3 @@
 *.csv
 baerer
+*.bak
diff --git a/dhl.sh b/dhl.sh
index 6ea2c36..620a737 100755
--- a/dhl.sh
+++ b/dhl.sh
@@ -1,13 +1,24 @@
 #!/usr/bin/env sh
+dependencies="sed test cp libreoffice"
+assert_tools () {
+	while test $# -gt 0; do
+		which $1  >/dev/null 2>/dev/null || {
+			echo "tool missing: "$1
+			exit 2
+		}
+		shift
+	done
+}
+assert_tools ${dependencies}
 
 # target filename
-tn="dhl-ds-versendung.csv"
+tn="dhl-cp1252.csv"
 
 # todo: download file (its UTF-8 encoded)
 
-fn="dhl-ds-versendung-3.csv"
+fn="dhl.csv"
 # backup
-cp -f "$fn" "dhl.csv"
+cp -f "${fn}" "${fn}.bak"
 
 # replace special spaces and hyphens
 sed -i -e 's/[\u00A0\u202F[:space:]]+/ /g' -e 's/[\u2010\u2011\u2012\u2013\u002D]+/-/g' -e 's/[\u00AD\uFEFF]+//g' "$fn"
@@ -77,3 +88,19 @@ testiso3 () {
 	testcmp "" "iso3countrycode" "Simbabwe" # should fail
 }
 #~ testiso3
+
+convcsv () {
+	if test $# -eq 0 ; then echo "argument missing."; exit 1; fi
+	if test -e $1 ; then echo "file already exists."; exit 2; fi 
+# charset windows codepage 1252
+# doc for headless conversion options https://wiki.openoffice.org/wiki/Documentation/DevGuide/Spreadsheets/Filter_Options
+	if="${1}"
+	of=$(echo "${if}"|rev|cut -d. -f2-|rev)".ods"
+	libreoffice --headless --convert-to ods --infilter=CSV:59,34,76,1 "${if}" && {
+		mv "${if}" "${if}.bak"
+		libreoffice --headless --infilter=CSV:59,34,76,1 --convert-to csv:"Text - txt - csv (StarCalc)":"44,34,1,0,1,,0" "${tf}" && {
+			rm "${tf}" "${if}.bak"
+		}
+	}
+}
+convcsv "$fn"