parsing

prod_data_parse_scripts/parse_uc_osnova.sh

@@ -11,7 +11,7 @@ input=$1
 output=${2:--}
 
 process_csv() {
-    awk -F ',' '
+    awk -F '\",\"' '
     function clean_account(value, quoted) {
         quoted = substr(value, 1, 1) == "\"" &&
                  substr(value, length(value), 1) == "\""
@@ -44,7 +44,7 @@ process_csv() {
     }
 
     BEGIN {
-        OFS = ","
+        OFS = FS
     }
 
     {
@@ -61,7 +61,7 @@ process_csv() {
         key = key_value(cleaned)
 
         if (key == "") {
-            cleaned = "\"000\""
+            cleaned = "000"
             key = "000"
         }
 

prod_data_parse_scripts/parse_uctovnictvo.sh

@@ -0,0 +1,75 @@
+#!/bin/sh
+
+set -eu
+
+if [ "$#" -lt 1 ] || [ "$#" -gt 2 ]; then
+    printf 'Usage: %s INPUT.csv [OUTPUT.csv]\n' "$0" >&2
+    exit 2
+fi
+
+input=$1
+output=${2:--}
+
+process_csv() {
+    awk -F '\",\"' '
+    function clean_account(value) {
+        if (length(value) > 0 && substr(value, length(value), 1) !~ /[0-9]/) {
+            while (length(value) > 0 && substr(value, length(value), 1) !~ /[0-9]/) {
+                value = substr(value, 1, length(value) - 1)
+            }
+        }
+
+        return value
+    }
+
+    function zero_amount_to_empty(value) {
+        if (value == "0.00") {
+            return ""
+        }
+
+        return value
+    }
+
+    BEGIN {
+        OFS = FS
+    }
+
+    NR == 1 {
+        print
+        next
+    }
+
+    {
+        cleaned = clean_account($10)
+
+        if (cleaned == "") {
+            cleaned = "000"
+        }
+
+        key = cleaned
+        key_count[key]++
+        if (!(key in key_order)) {
+            key_order[key] = ++key_total
+        }
+
+        $10 = cleaned
+        $11 = zero_amount_to_empty($11)
+        $12 = zero_amount_to_empty($12)
+        print
+    }
+
+    END {
+        for (key in key_order) {
+            if (key_count[key] > 1) {
+                printf "Non-unique column 10 key: %s (%d rows)\n", key, key_count[key] > "/dev/stderr"
+            }
+        }
+    }
+' "$input"
+}
+
+if [ "$output" = "-" ]; then
+    process_csv
+else
+    process_csv > "$output"
+fi