如何用sed替換不帶空格的信件？

Question

我需要刪除一切，這是不是一個字母，大寫或小寫，從一個文件中，並用空格代替它，例如：

The bear ate 3 snakes, then ate 50% of the fish from the river. 

這將成爲：

The bear ate snakes then ate  of the fish from the river 

• 有時文件包含不尋常的字符它被保存爲UTF-8。

如何用空格替換任何非字母？

Answer 1

$ echo "The bear ate 3 snakes, then ate 50% of the fish from the river." | sed "s/[^a-zA-Z]/ /g" 
The bear ate snakes then ate  of the fish from the river 

Answer 2

嘗試：

sed 's/[^A-Za-z]/ /g;' myfile.txt 

Answer 3

如果你想爲Unicode字母(as mentioned in your question)支持那麼這個perl的命令，以取代所有unicode non-letters：

echo $line | perl -pe 's/[^\p{L}\s]+/ /g;' 

參考：http://www.regular-expressions.info/unicode.html

Answer 4

這可能會實現爲你：

echo 'The bear ate 3 snakes, then ate 50% of the fish from the river.' | 
tr -c '[:alpha:]' ' ' 
The bear ate snakes then ate  of the fish from the river 

或：

echo 'The bear ate 3 snakes, then ate 50% of the fish from the river.' | 
sed 's/[^[:alpha:]]/ /g' 
The bear ate snakes then ate  of the fish from the river