如何在 CSVFormat 中避免逗号前的反斜杠

Question

我在 java 中使用 CSVFormat 创建一个 CSV 文件，我在标题和值中面临的问题是每当字符串很长并且有逗号时，api 总是在逗号之前插入一个 \。结果，标头没有正确形成，并且 csv 文件中的值将下一个单元格用于 .我正在发布我所做的代码

   try (CSVPrinter csvPrinter = new CSVPrinter(out,
            CSVFormat.DEFAULT.withHeader("\""+SampleEnum.MY_NAME.getHeader()+"\"", "\""+SampleEnum.MY_TITLE.getHeader()+"\"",
                    "\""+SampleEnum.MY_ID.getHeader()+"\"", "\""+SampleEnum.MY_NUMBER.getHeader()+"\"", "\""+SampleEnum.MY_EXTERNAL_KEY.getHeader()+"\"",
                    "\""+SampleEnum.DATE.getHeader()+"\"","\""+SampleEnum.MY_ACTION.getHeader()+"\"",
                    "\"\"\""+SampleEnum.MY__DEFI.getHeader()+"\"\"\"", SampleEnum.MY_ACTION.getHeader(),
                    SampleEnum.CCHK.getHeader(), SampleEnum.DISTANCE_FROM_LOCATION.getHeader(),
                    SampleEnum.TCOE.getHeader(), SampleEnum.HGTR.getHeader(),SampleEnum._BLANK.getHeader(),
                    SampleEnum.LOCATION_MAP.getHeader(), SampleEnum.SUBMISSION_ID.getHeader())                      
                    .withDelimiter(',').withEscape('\\').withQuote('"').withTrim().withQuoteMode(QuoteMode.MINIMAL)
    )) {
        sampleModel.forEach(sf -> {
            try {
                csvPrinter.printRecord(sf.getMyName(),
                        sf.getMyTitle(),
                        sf.getMyID(),
                        sf.getMyNo(),

所以现在的问题是我得到这样的输出

"\"Name:\"","\"Title\"","\"ID #:\"","\"Store #:\"","\"Store #: External Key\"","\"Date:\"","\"\"\"It's performance  issue in detail to include dates,times, circumstances, etc.\"\"\""

我在每个逗号之前得到 \，当它出现在值中时，文本的下一部分将转移到下一个单元格。

我需要的输出是

"Name:","Title:","Employee ID #:","Store #:","Store #: CurrierKey","Date:","Stage of Disciplinary Action:","""Describe your view about the company, times, circumstances, etc.""",

我正在尝试 https://commons.apache.org/proper/commons-csv/jacoco/org.apache.commons.csv/CSVFormat.java.html 这个链接，但我无法理解修复。请帮忙。

Answer 1

发生这种情况是因为您使用的是 QuoteMode.NONE，它具有以下 Javadoc：

从不引用字段。当分隔符出现在数据中时，打印机会在其前面加上转义字符。如果未设置转义字符，格式验证会抛出异常。

您可以使用QuoteMode.MINIMAL 仅引用包含特殊字符的字段（例如字段分隔符、引号字符或行分隔符字符串的字符）。

---

我建议您使用CSVFormat.DEFAULT，然后如果您不能使用其他格式之一，请自行配置所有内容。检查反斜杠 (\) 是否真的是您的用例的正确转义字符。通常它是双引号 (")。此外，您可能希望从标题定义中删除所有双引号，因为它们会根据您的配置自动添加（如果需要）。

StringBuilder out = new StringBuilder();
try (CSVPrinter csvPrinter = new CSVPrinter(out,
        CSVFormat.DEFAULT
                .withHeader("AAAA", "BB\"BB", "CC,CC", "DD'DD")
                .withDelimiter(',')
                .withEscape('\\') // <- maybe you want '"' instead
                .withQuote('"').withRecordSeparator('\n').withTrim()
                .withQuoteMode(QuoteMode.MINIMAL)
)) {
    csvPrinter.printRecord("WWWW", "XX\"XX", "YY,YY", "ZZ'ZZ");
}
System.out.println(out);

AAAA,"BB\"BB","CC,CC",DD'DD
WWWW,"XX\"XX","YY,YY",ZZ'ZZ

---

在您的edit 之后，您似乎希望所有字段都用双引号作为转义字符引用。因此，您可以像这样使用QuoteMode.ALL 和.withEscape('"')：

StringBuilder out = new StringBuilder();
try (CSVPrinter csvPrinter = new CSVPrinter(out,
        CSVFormat.DEFAULT
                .withHeader("AAAA", "BB\"BB", "CC,CC", "\"DD\"", "1")
                .withDelimiter(',')
                .withEscape('"')
                .withQuote('"').withRecordSeparator('\n').withTrim()
                .withQuoteMode(QuoteMode.ALL)
)) {
    csvPrinter.printRecord("WWWW", "XX\"XX", "YY,YY", "\"DD\"", "2");
}
System.out.println(out);

"AAAA","BB""BB","CC,CC","""DD""","1"
"WWWW","XX""XX","YY,YY","""DD""","2"

---

在您的comment 中，您声明您只需要在需要时使用双引号，并且仅在一个字段中使用三引号。然后，您可以按照第一个示例中的建议使用QuoteMode.MINIMAL 和.withEscape('"')。当您用双引号将该字段的输入括起来时，会生成三引号（一次是因为有一个特殊字符并且需要引用该字段，第二个是因为您添加了显式 " 而第三个是为了逃避你的明确报价）。

StringBuilder out = new StringBuilder();
try (CSVPrinter csvPrinter = new CSVPrinter(out,
        CSVFormat.DEFAULT
                .withHeader("AAAA", "BB\"BB", "CC,CC", "\"DD\"", "1")
                .withDelimiter(',')
                .withEscape('"')
                .withQuote('"').withRecordSeparator('\n').withTrim()
                .withQuoteMode(QuoteMode.MINIMAL)
)) {
    csvPrinter.printRecord("WWWW", "XX\"XX", "YY,YY", "\"DD\"", "2");
}
System.out.println(out);

AAAA,"BB""BB","CC,CC","""DD""",1
WWWW,"XX""XX","YY,YY","""DD""",2

---

根据chat，您希望完全控制标题是否包含引号。没有QuoteMode 和转义字符的组合可以提供所需的结果。因此，我建议您手动构建标题：

StringBuilder out = new StringBuilder();
try (CSVPrinter csvPrinter = new CSVPrinter(out,
        CSVFormat.DEFAULT
                .withDelimiter(',').withEscape('"')
                .withQuote('"').withRecordSeparator('\n').withTrim()
                .withQuoteMode(QuoteMode.MINIMAL))
) {
    out.append(String.join(",", "\"AAAA\"", "\"BBBB\"", "\"CC,CC\"", "\"\"\"DD\"\"\"", "1"));
    out.append("\n");
    csvPrinter.printRecord("WWWW", "XX\"XX", "YY,YY", "\"DD\"", "2");
}
System.out.println(out);

"AAAA","BBBB","CC,CC","""DD""",1
WWWW,"XX""XX","YY,YY","""DD""",2