awk:超出程序限制:最大字段数 size=32767

Question

当我在 ubuntu 14.04 中运行我的 shell 脚本时，我收到如下错误

awk: program limit exceeded: maximum number of fields size=32767
    FILENAME="ensemble_features/Training_BOOKS_POS_Bigram_with_stemming_BOOLEAN_FVT.csv" FNR=1 NR=1
cut: invalid byte, character or field list
Try 'cut --help' for more information.
-1
cut: invalid byte, character or field list
Try 'cut --help' for more information.
6656
user@user-Lenovo-IdeaPad-Z410:~/Thesis/BOOKS$ bash Training_POS_Uni_Bi.sh
awk: program limit exceeded: maximum number of fields size=32767
    FILENAME="ensemble_features/Training_BOOKS_POS_Bigram_with_stemming_BOOLEAN_FVT.csv" FNR=1 NR=1
cut: invalid byte, character or field list
Try 'cut --help' for more information.
-1
cut: invalid byte, character or field list
Try 'cut --help' for more information.
6656

我在下面添加我的脚本

cd /home/user/Thesis/BOOKS/Features/Training/POSITIVE/
fname="ensemble_features"
mkdir $fname

cp /home/user/Thesis/BOOKS/Features/Training/POSITIVE/Training_BOOKS_POS_unigram_FVT_with_stemming_BOOLEAN.csv ensemble_features/
cp /home/user/Thesis/BOOKS/Features/Training/POSITIVE/Training_BOOKS_POS_Bigram_with_stemming_BOOLEAN_FVT.csv ensemble_features/


mkdir "proces"
cnt=0
for file in $fname/*
do
    #Number of columns
    num=`awk 'BEGIN {FS=",";c=0};{if (c==0 ){print NF; c=1}}END{}' $file`
    if [[ cnt -eq 0 ]];then
        cut -d, -f $num $file >class.csv
        cnt=1;
    fi
    num=$((num-1))
    echo $num
    nfname=`basename $file`

    #Cut the columns
    cut -d',' -f1-$num $file > proces/cutlast$nfname
done
#Paste multiple csv
paste -d',' proces/* > comb.csv
paste -d, comb.csv class.csv > Training_BOOKS_Unigram_Bigram_POS_Ensemble_Features_BOOLEAN.csv
rm comb.csv
rm class.csv
rm -r proces
rm -r ensemble_features

我的输入文件分别包含 38453 列和 6656 列。有人可以帮我纠正这个错误吗？

Answer 1

在 ubuntu 上，awk 是一些 awk 变体的软链接(soft link)，现在默认情况下它是 mawk。尝试安装 gawk。 gawk 对记录中的字段数没有限制。

顺便说一句，python 可能是一个更好的长期解决方案，如果你有时间学习它。