正则表达式替换字符答案

【问题标题】：Regexp replacement character正则表达式替换字符
【发布时间】：2021-08-08 13:29:04
【问题描述】：

我在 Go 中创建了一个 CSV 文件，我必须在每一列中添加引号（“），我添加了这些，但是这一次，CSV 编程在 comment 中添加了额外的（双）引号 em> 列（如果列中有逗号（,））

我的 CSV

comment_ID","post_ID","product_SKU","comment_author","author_mail","author_location","date","comment"
"100","60574","VID17","Jordi","","","2021-06-02 16:20:30",""My son likes this video, good job""
"101","60574","VID18","Scarlett,"","","2020-12-29 23:06:32",""I don't like this video, it may be better""

我需要这样的CSV（评论栏中没有双引号）

comment_ID","post_ID","product_SKU","comment_author","author_mail","author_location","date","comment"
"100","60574","VID17","Jordi","","","2021-06-02 16:20:30","My son likes this video, good job"
"101","60574","VID18","Scarlett,"","","2020-12-29 23:06:32","I don't like this video, it may be better"

我的 Golang 代码

RegContent := regexp.MustCompile(`",""[A-Za-z0-9]`)
newRegexp := RegContent.ReplaceAllString(CSV_Contents, `","`)
fmt.Println("PLAY: ", newRegexp)
err = ioutil.WriteFile(path, []byte(newRegexp), 0)
if err != nil {
    fmt.Println("error: ", err)
}

输出

"son likes this video, good job" //(Missing My)
"don't like this video, it may be better" //(Missing I)

【问题讨论】：

您当前和预期的 CSV 文件均无效； oeader 行中的第一列缺少开头的双引号。
@Wiktor Stribiże，不，这也替换了 "Scarlett,"","" => "Scarlett,"," 和 "乔迪","","" => "乔迪",", "
@Axifive 我明白了，空字段会受到影响。所以，剩下的唯一问题就是评论栏里面有没有""，怎么处理。
您的 CSV 在 ,"Scarlett,"" 上也有问题，它应该在 Scarlett 之后包含一个引号。您是如何生成该 CSV 的？真的很糟糕；如果可能，请重新生成您的 csv，而不是尝试解决此问题。除非你只是在练习。但是我不应该帮助你太多。我只是建议远离正则表达式。

标签： regex string csv go

【解决方案1】：

您可以在捕获外引号之间的所有内容时匹配最后一列，并在 ReplaceAllString 的替换参数中使用反向引用来恢复该部分：

package main

import (
    "fmt"
    "regexp"
)

func main() {
        CSV_Contents := `
comment_ID","post_ID","product_SKU","comment_author","author_mail","author_location","date","comment"
"100","60574","VID17","Jordi","","","2021-06-02 16:20:30",""My son likes this video, good job""
"101","60574","VID18","Scarlett,"","","2020-12-29 23:06:32",""I don't like this video, it may be better""
`   
    RegContent := regexp.MustCompile(`(?m),"("[^"]*(?:""[^"]*)*")"$`)
    result := RegContent.ReplaceAllString(CSV_Contents, `,$1`)
    fmt.Println(result)
}

见Go demo，输出：

comment_ID","post_ID","product_SKU","comment_author","author_mail","author_location","date","comment"
"100","60574","VID17","Jordi","","","2021-06-02 16:20:30","My son likes this video, good job"
"101","60574","VID18","Scarlett,"","","2020-12-29 23:06:32","I don't like this video, it may be better"

请参阅regex demo。详情：

(?m) - 开启多行模式，$ 将匹配行尾
," - 逗号和"
("[^"]*(?:""[^"]*)*") - 第 1 组 ($1)："，然后是除 " 之外的任何零个或多个字符，然后是零个或多个 "" 序列（如果注释列中有转义引号，它们将保持不变），然后是零个或多个非" 字符，然后
"$ - 一行末尾的 "。

【讨论】：

它在终端上工作，但如果我在 CSV 文件上运行它，它就不起作用。感谢您的努力
@Melisa 这意味着您的内容与您在问题中发布的内容不同。可能行尾有尾随空格，regexp.MustCompile(`(?m),"("[^"]*(?:""[^"]*)*")"\s*$`) 会起作用，但如果没有确切的文件，则无法保证。

【解决方案2】：

您可以使用 ReplaceAllStringFunc() 获得描述的行为

f := func(s string) string {
   return strings.ReplaceAll(s, `""`, `"`)
}
RegContent := regexp.MustCompile(`",""[^,].+""`)
newRegexp := RegContent.ReplaceAllStringFunc(CSV_Contents, f)
fmt.Println("PLAY: ", newRegexp)

https://play.golang.org/p/1NqTyN1hs1J

还有ReplaceAllString()的替代方案：

RegContent := regexp.MustCompile(`,""([^,].+)""`)
newRegexp := RegContent.ReplaceAllString(CSV_Contents, `,"$1"`)
fmt.Println("PLAY: ", newRegexp)

https://play.golang.org/p/tY8zGWTbLLB

【讨论】：

替代方法（replaceAllString）直接起作用！