Question

我有这些数据：

$ head test
"Rec Open Date","MSISDN","IMEI"
"2015-01-04",100,200
"2015-01-04",100,200
"2015-01-04",100,200
"2015-01-04",100,200
"2015-01-04",100,200
"2015-01-04",100,200
"2015-01-04",100,200
"2015-01-04",100,200
"2015-01-04",100,200

我可以使用此命令更改第一列的值：

$ awk -F, 'NR>1{$1="2015-01-05"}1' OFS=, test > tmpfile && mv tmpfile test

用这个命令我松开双引号，我想保留双引号。可以修改此命令以实现此目的吗？

$ head test
"Rec Open Date","MSISDN","IMEI"
2015-01-05,100,200
2015-01-05,100,200
2015-01-05,100,200
2015-01-05,100,200
2015-01-05,100,200
2015-01-05,100,200
2015-01-05,100,200
2015-01-05,100,200
2015-01-05,100,200

Answer 1

来自GNU Awk用户指南的

;

\＆＃34;文字双引号（仅限字符串常量）。这个当你想写一个字符串常量时，使用sequence 包含双引号。因为字符串是由double分隔的引号，你需要转义作为字符串一部分的引用为了告诉awk继续处理字符串的其余部分。

$ awk -F, 'NR>1{$1="\"2015-01-05\""}1' OFS=, test > tmpfile && mv tmpfile test

$ head test
"Rec Open Date","MSISDN","IMEI"
"2015-01-05",100,200
"2015-01-05",100,200
"2015-01-05",100,200
"2015-01-05",100,200
"2015-01-05",100,200
"2015-01-05",100,200
"2015-01-05",100,200
"2015-01-05",100,200
"2015-01-05",100,200

Answer 2

并非您“失去”它们，而是$1最初为"something"并且您将其替换为another thing。由于此another thing是一个字符串，因此您可以使用双引号来记录它。因此，为了获得双引号，您需要在进行替换时指明。

这可以更清洁：

awk -v new='"2015-01-05"' 'BEGIN {FS=OFS=","} NR>1{$1=new}1' file
    ^^^^^^^^^^^^^^^^^^^^^                              ^
         |---------------------------------------------|
    we provide the var ............... and we replace with it
      surrounded by "

awk +更改列值并保留双引号

2 个答案: