diff --git a/lib/ulib/awk b/lib/ulib/awk
index c242e6b..2a289b0 100644
--- a/lib/ulib/awk
+++ b/lib/ulib/awk
@@ -841,7 +841,7 @@ function lawkcsv() {
         reset_fields=
     fi
 
-    awkrun -f \
+    lawkrun -f \
         skip_lines:int="$skip" parse_headers:int="$parse_headers" \
         autosep:int="$autosep" colsep="$colsep" qchar="$qchar" echar="$echar" \
         show_headers:int="$show_headers" reset_fields:int="$reset_fields" \
@@ -978,7 +978,7 @@ function lgrepcsv() {
     headers=("$@")
     if [ -z "${headers[*]}" ]; then
         parse_headers=1
-        array_from_lines headers "$(awkrun -f lskip:int="$skip" 'NR <= lskip { next }
+        array_from_lines headers "$(lawkrun -f lskip:int="$skip" 'NR <= lskip { next }
 {
   count = array_parsecsv(__fields, $0)
   for (i = 1; i <= count; i++) {
@@ -1077,7 +1077,7 @@ function lawkfsv2csv() {
     done
     shift
 
-    awkrun -f \
+    lawkrun -f \
         skip_lines:int="$skip" trim_values:int="$rtrim" show_headers:int="$show_headers" \
         headers[@] starts[@] sizes[@] \
         "$__AWKCSV_FUNCTIONS"'
@@ -1109,6 +1109,57 @@ function awkfsv2csv() { LANG=C lawkfsv2csv "$@"; }
 
 ################################################################################
 
+__AWKCSV2FSV_HELP="\
+Transformer un flux csv en fsv (colonnes à largeurs fixes)
+
+Chaque argument doit être de la forme field:size"
+
+function lawkcsv2fsv() {
+    local -a args fieldsizes
+    if parse_opts @ args -- "$@"; then
+        set -- "${args[@]}"
+    else
+        eerror "$args"
+        return 1
+    fi
+
+    local -a fields sizes
+    local fieldsize field size
+    while [ $# -gt 0 -a "$1" != "--" ]; do
+        fieldsize="$1"
+        shift
+
+        splitpair "$fieldsize" field size
+        [ -n "$field" ] || {
+            eerror "field est requis"
+            return 1
+        }
+        [ -n "$size" ] || size=1
+        array_add fields "$field"
+        array_add sizes "$size"
+    done
+    shift
+
+    lawkcsv -v fields[@] -v sizes[@] -a '
+{
+  line = ""
+  for (i = 1; i <= fields_count; i++) {
+    size = sizes[i] + 0
+    value = get(fields[i])
+    while (length(value) < size) value = value " "
+    if (length(value) > size) value = substr(value, 1, size)
+    line = line value
+  }
+  print line
+}
+' -- "$@"
+}
+
+function cawkcsv2fsv() { LANG=C lawkcsv2fsv "$@"; }
+function awkcsv2fsv() { LANG=C lawkcsv2fsv "$@"; }
+
+################################################################################
+
 __MERGECSV_HELP="\
 Fusionner deux fichiers csv en faisant la correspondance sur la valeur d'un
 champ, qui est la clé
@@ -1264,14 +1315,14 @@ function lmergecsv() {
 
     [ "$parse_headers" == "auto" ] && parse_headers=1
     if [ -n "$parse_headers" ]; then
-        [ -n "$lheaders" ] || lheaders="$(<"$lfile" awkrun lskip:int="$lskip" 'NR <= lskip { next } { print; exit }')"
-        [ -n "$rheaders" ] || rheaders="$(<"$rfile" awkrun lskip:int="$lskip" 'NR <= lskip { next } { print; exit }')"
+        [ -n "$lheaders" ] || lheaders="$(<"$lfile" lawkrun lskip:int="$lskip" 'NR <= lskip { next } { print; exit }')"
+        [ -n "$rheaders" ] || rheaders="$(<"$rfile" lawkrun lskip:int="$lskip" 'NR <= lskip { next } { print; exit }')"
     fi
 
     # faire le fichier de travail pour lfile
     local tmpleft
     ac_set_tmpfile tmpleft "" __mergecsv_left "" __MERGECSV_DEBUG
-    <"$lfile" awkrun -f \
+    <"$lfile" lawkrun -f \
         padding="$padding" padlen:int="$padlen" \
         parse_headers:int="$parse_headers" ignore_case:int="$ignore_case" \
         lskip:int="$lskip" lkey="$lkey" \
@@ -1295,7 +1346,7 @@ parse_headers && do_once("parse-headers") {
     # faire le fichier de travail pour rfile
     local tmpright
     ac_set_tmpfile tmpright "" __mergecsv_right "" __MERGECSV_DEBUG
-    <"$rfile" awkrun -f \
+    <"$rfile" lawkrun -f \
         padding="$padding" padlen:int="$padlen" \
         parse_headers:int="$parse_headers" ignore_case:int="$ignore_case" \
         rskip:int="$rskip" rkey="$rkey" \
@@ -1409,7 +1460,7 @@ parse_headers && do_once("parse-headers") {
     ac_set_tmpfile tmpmerged "" __mergecsv_merged "" __MERGECSV_DEBUG
     csort -s -k 1,$(($padlen + 1)) "$tmpleft" "$tmpright" >"$tmpmerged"
 
-    <"$tmpmerged" awkrun -f \
+    <"$tmpmerged" lawkrun -f \
         padlen:int="$padlen" \
         parse_headers:int="$parse_headers" ignore_case:int="$ignore_case" \
         lheaderscsv="$lheaders" lkey="$lkey" lprefix="$lprefix" \
@@ -1700,14 +1751,14 @@ function lsortcsv() {
             cat >"$tmpinput"
             input="$tmpinput"
         fi
-        headers="$(<"$input" awkrun skip:int="$skip" 'NR <= skip { next } { print; exit }')"
+        headers="$(<"$input" lawkrun skip:int="$skip" 'NR <= skip { next } { print; exit }')"
     fi
 
     # faire le fichier de travail
     local tmpinput
     ac_set_tmpfile tmpinput "" __sortcsv_input "" __SORTCSV_DEBUG
     array_add tmpfiles "$tmpinput"
-    <"$input" >"$tmpinput" awkrun -f \
+    <"$input" >"$tmpinput" lawkrun -f \
         padding="$padding" padlen:int="$padlen" \
         skip:int="$skip" parse_headers:int="$parse_headers" \
         key="$key" \
@@ -1740,7 +1791,7 @@ parse_headers && do_once("parse-headers") {
 
     # résultat
     [ -n "$output" ] || output=/dev/stdout
-    stdredir "$tmpsorted" "$output" "" awkrun -f \
+    stdredir "$tmpsorted" "$output" "" lawkrun -f \
         padlen:int="$padlen" \
         headerscsv="$headers" show_headers:int="$show_headers" \
         '
@@ -2060,7 +2111,7 @@ function lprintcsv() {
     [ -n "$output" ] || output=/dev/stdout
     values=("$@")
 
-    stdredir "" ">>$output" "" awkrun -f fields[@] show_headers:int="$show_headers" values[@] '
+    stdredir "" ">>$output" "" lawkrun -f fields[@] show_headers:int="$show_headers" values[@] '
 BEGIN {
   if (show_headers) array_printcsv(fields)
   if (fields_count > 0) count = fields_count
diff --git a/uawk b/uawk
index bd9716a..c59351a 100755
--- a/uawk
+++ b/uawk
@@ -1,7 +1,7 @@
 #!/bin/bash
 # -*- coding: utf-8 mode: sh -*- vim:sw=4:sts=4:et:ai:si:sta:fenc=utf-8
 
-TOOLS=(awkrun awkcsv grepcsv awkfsv2csv mergecsv sortcsv dumpcsv printcsv)
+TOOLS=(awkrun awkcsv grepcsv awkfsv2csv awkcsv2fsv mergecsv sortcsv dumpcsv printcsv)
 if [ "$#" -eq 1 -a "$1" == --nutools-makelinks ]; then
     # créer les liens
     scriptname="$(basename "$0")"
@@ -49,6 +49,7 @@ run) tool=awkrun;;
 csv) tool=awkcsv;;
 grep) tool=grepcsv;;
 fsv2csv|fsv) tool=awkfsv2csv;;
+csv2fsv) tool=awkcsv2fsv;;
 merge) tool=mergecsv;;
 sort) tool=sortcsv;;
 dump) tool=dumpcsv;;
@@ -95,12 +96,23 @@ elif [ "$tool" == "awkfsv2csv" ]; then
 ${Clang:+note: ce script ne tient pas compte de la locale: il travaille avec LANG=C
 }
 USAGE
-    $scriptname [options] -- [header:sizes...] [-- inputfiles...]
+    $scriptname [options] -- [header:size,...] [-- inputfiles...]
 
 $__AWKFSV2CSV_HELP"
 
     "${prefix}awkfsv2csv" "$@"
 
+elif [ "$tool" == "awkcsv2fsv" ]; then
+    [ $# -eq 1 -a "$1" == --help ] && exit_with uecho "$scriptname: wrapper autour de awk pour transformer un fichier à colonnes fixes en fichier csv
+${Clang:+note: ce script ne tient pas compte de la locale: il travaille avec LANG=C
+}
+USAGE
+    $scriptname [options] -- [field:size,...] [-- inputfiles...]
+
+$__AWKCSV2FSV_HELP"
+
+    "${prefix}awkcsv2fsv" "$@"
+
 elif [ "$tool" == "mergecsv" ]; then
     [ $# -eq 1 -a "$1" == --help ] && exit_with uecho "$scriptname: fusionner deux fichiers csv sur un champ commun
 ${Clang:+note: ce script ne tient pas compte de la locale: il travaille avec LANG=C