obrobka.sh

#!/usr/bin/env bash

# test 
# head -c 10M 17.1-EX_XML_EDR_UO_FULL_29.05.2020.xml > test.xml
# XML_FILES=test.xml

XML_FILES=17*.xml

for f in $XML_FILES
    do
        echo $f
        split -d  -a 3 -b 100M $f _s
        SPLITTED_FILES=_s*
        rm $f
        for sf in $SPLITTED_FILES
            do
                echo $sf
                sed -i 's/\r//g' $sf
                # cause error
                # recode CP1251..UTF-8 $sf
                iconv -f cp1251 -t utf8 "$sf">"$sf.new" && mv -f "$sf.new" "$sf"
                # if chunk is the first
                if [[ $sf =~ ^_s000$ ]]; then
                    sed -i 's/="windows-1251"/="utf-8"/g' $sf
                fi
            done
        cat $SPLITTED_FILES > 'p_'$f
        rm $SPLITTED_FILES
    done