Table of Contents for
Regular Expressions Cookbook, 2nd Edition

Version ebook / Retour

Cover image for bash Cookbook, 2nd Edition Regular Expressions Cookbook, 2nd Edition by Steven Levithan Published by O'Reilly Media, Inc., 2012
  1. Cover
  2. Regular Expressions Cookbook
  3. Preface
  4. Caught in the Snarls of Different Versions
  5. Intended Audience
  6. Technology Covered
  7. Organization of This Book
  8. Conventions Used in This Book
  9. Using Code Examples
  10. Safari® Books Online
  11. How to Contact Us
  12. Acknowledgments
  13. 1. Introduction to Regular Expressions
  14. Regular Expressions Defined
  15. Search and Replace with Regular Expressions
  16. Tools for Working with Regular Expressions
  17. 2. Basic Regular Expression Skills
  18. 2.1. Match Literal Text
  19. 2.2. Match Nonprintable Characters
  20. 2.3. Match One of Many Characters
  21. 2.4. Match Any Character
  22. 2.5. Match Something at the Start and/or the End of a Line
  23. 2.6. Match Whole Words
  24. 2.7. Unicode Code Points, Categories, Blocks, and Scripts
  25. 2.8. Match One of Several Alternatives
  26. 2.9. Group and Capture Parts of the Match
  27. 2.10. Match Previously Matched Text Again
  28. 2.11. Capture and Name Parts of the Match
  29. 2.12. Repeat Part of the Regex a Certain Number of Times
  30. 2.13. Choose Minimal or Maximal Repetition
  31. 2.14. Eliminate Needless Backtracking
  32. 2.15. Prevent Runaway Repetition
  33. 2.16. Test for a Match Without Adding It to the Overall Match
  34. 2.17. Match One of Two Alternatives Based on a Condition
  35. 2.18. Add Comments to a Regular Expression
  36. 2.19. Insert Literal Text into the Replacement Text
  37. 2.20. Insert the Regex Match into the Replacement Text
  38. 2.21. Insert Part of the Regex Match into the Replacement Text
  39. 2.22. Insert Match Context into the Replacement Text
  40. 3. Programming with Regular Expressions
  41. Programming Languages and Regex Flavors
  42. 3.1. Literal Regular Expressions in Source Code
  43. 3.2. Import the Regular Expression Library
  44. 3.3. Create Regular Expression Objects
  45. 3.4. Set Regular Expression Options
  46. 3.5. Test If a Match Can Be Found Within a Subject String
  47. 3.6. Test Whether a Regex Matches the Subject String Entirely
  48. 3.7. Retrieve the Matched Text
  49. 3.8. Determine the Position and Length of the Match
  50. 3.9. Retrieve Part of the Matched Text
  51. 3.10. Retrieve a List of All Matches
  52. 3.11. Iterate over All Matches
  53. 3.12. Validate Matches in Procedural Code
  54. 3.13. Find a Match Within Another Match
  55. 3.14. Replace All Matches
  56. 3.15. Replace Matches Reusing Parts of the Match
  57. 3.16. Replace Matches with Replacements Generated in Code
  58. 3.17. Replace All Matches Within the Matches of Another Regex
  59. 3.18. Replace All Matches Between the Matches of Another Regex
  60. 3.19. Split a String
  61. 3.20. Split a String, Keeping the Regex Matches
  62. 3.21. Search Line by Line
  63. Construct a Parser
  64. 4. Validation and Formatting
  65. 4.1. Validate Email Addresses
  66. 4.2. Validate and Format North American Phone Numbers
  67. 4.3. Validate International Phone Numbers
  68. 4.4. Validate Traditional Date Formats
  69. 4.5. Validate Traditional Date Formats, Excluding Invalid Dates
  70. 4.6. Validate Traditional Time Formats
  71. 4.7. Validate ISO 8601 Dates and Times
  72. 4.8. Limit Input to Alphanumeric Characters
  73. 4.9. Limit the Length of Text
  74. 4.10. Limit the Number of Lines in Text
  75. 4.11. Validate Affirmative Responses
  76. 4.12. Validate Social Security Numbers
  77. 4.13. Validate ISBNs
  78. 4.14. Validate ZIP Codes
  79. 4.15. Validate Canadian Postal Codes
  80. 4.16. Validate U.K. Postcodes
  81. 4.17. Find Addresses with Post Office Boxes
  82. 4.18. Reformat Names From “FirstName LastName” to “LastName, FirstName”
  83. 4.19. Validate Password Complexity
  84. 4.20. Validate Credit Card Numbers
  85. 4.21. European VAT Numbers
  86. 5. Words, Lines, and Special Characters
  87. 5.1. Find a Specific Word
  88. 5.2. Find Any of Multiple Words
  89. 5.3. Find Similar Words
  90. 5.4. Find All Except a Specific Word
  91. 5.5. Find Any Word Not Followed by a Specific Word
  92. 5.6. Find Any Word Not Preceded by a Specific Word
  93. 5.7. Find Words Near Each Other
  94. 5.8. Find Repeated Words
  95. 5.9. Remove Duplicate Lines
  96. 5.10. Match Complete Lines That Contain a Word
  97. 5.11. Match Complete Lines That Do Not Contain a Word
  98. 5.12. Trim Leading and Trailing Whitespace
  99. 5.13. Replace Repeated Whitespace with a Single Space
  100. 5.14. Escape Regular Expression Metacharacters
  101. 6. Numbers
  102. 6.1. Integer Numbers
  103. 6.2. Hexadecimal Numbers
  104. 6.3. Binary Numbers
  105. 6.4. Octal Numbers
  106. 6.5. Decimal Numbers
  107. 6.6. Strip Leading Zeros
  108. 6.7. Numbers Within a Certain Range
  109. 6.8. Hexadecimal Numbers Within a Certain Range
  110. 6.9. Integer Numbers with Separators
  111. 6.10. Floating-Point Numbers
  112. 6.11. Numbers with Thousand Separators
  113. 6.12. Add Thousand Separators to Numbers
  114. 6.13. Roman Numerals
  115. 7. Source Code and Log Files
  116. Keywords
  117. Identifiers
  118. Numeric Constants
  119. Operators
  120. Single-Line Comments
  121. Multiline Comments
  122. All Comments
  123. Strings
  124. Strings with Escapes
  125. Regex Literals
  126. Here Documents
  127. Common Log Format
  128. Combined Log Format
  129. Broken Links Reported in Web Logs
  130. 8. URLs, Paths, and Internet Addresses
  131. 8.1. Validating URLs
  132. 8.2. Finding URLs Within Full Text
  133. 8.3. Finding Quoted URLs in Full Text
  134. 8.4. Finding URLs with Parentheses in Full Text
  135. 8.5. Turn URLs into Links
  136. 8.6. Validating URNs
  137. 8.7. Validating Generic URLs
  138. 8.8. Extracting the Scheme from a URL
  139. 8.9. Extracting the User from a URL
  140. 8.10. Extracting the Host from a URL
  141. 8.11. Extracting the Port from a URL
  142. 8.12. Extracting the Path from a URL
  143. 8.13. Extracting the Query from a URL
  144. 8.14. Extracting the Fragment from a URL
  145. 8.15. Validating Domain Names
  146. 8.16. Matching IPv4 Addresses
  147. 8.17. Matching IPv6 Addresses
  148. 8.18. Validate Windows Paths
  149. 8.19. Split Windows Paths into Their Parts
  150. 8.20. Extract the Drive Letter from a Windows Path
  151. 8.21. Extract the Server and Share from a UNC Path
  152. 8.22. Extract the Folder from a Windows Path
  153. 8.23. Extract the Filename from a Windows Path
  154. 8.24. Extract the File Extension from a Windows Path
  155. 8.25. Strip Invalid Characters from Filenames
  156. 9. Markup and Data Formats
  157. Processing Markup and Data Formats with Regular Expressions
  158. 9.1. Find XML-Style Tags
  159. 9.2. Replace Tags with
  160. 9.3. Remove All XML-Style Tags Except and
  161. 9.4. Match XML Names
  162. 9.5. Convert Plain Text to HTML by Adding

    and
    Tags

  163. 9.6. Decode XML Entities
  164. 9.7. Find a Specific Attribute in XML-Style Tags
  165. 9.8. Add a cellspacing Attribute to Tags That Do Not Already Include It
  166. 9.9. Remove XML-Style Comments
  167. 9.10. Find Words Within XML-Style Comments
  168. 9.11. Change the Delimiter Used in CSV Files
  169. 9.12. Extract CSV Fields from a Specific Column
  170. 9.13. Match INI Section Headers
  171. 9.14. Match INI Section Blocks
  172. 9.15. Match INI Name-Value Pairs
  173. Index
  174. Index
  175. Index
  176. Index
  177. Index
  178. Index
  179. Index
  180. Index
  181. Index
  182. Index
  183. Index
  184. Index
  185. Index
  186. Index
  187. Index
  188. Index
  189. Index
  190. Index
  191. Index
  192. Index
  193. Index
  194. Index
  195. Index
  196. Index
  197. Index
  198. Index
  199. About the Authors
  200. Colophon
  201. Copyright

    202. 9.11. Change the Delimiter Used in CSV Files

    Problem

    You want to change all field-delimiting commas in a CSV file to tabs. Commas that occur within double-quoted values should be left alone.

    Solution

    The following regular expression matches an individual CSV field along with its preceding delimiter, if any. The preceding delimiter is usually a comma, but can also be an empty string (i.e., nothing) when matching the first field of the first record, or a line break when matching the first field of any subsequent record. Every time a match is found, the field itself, including the double quotes that may surround it, is captured to backreference 2, and its preceding delimiter is captured to backreference 1.

    Tip

    The regular expressions in this recipe are designed to work correctly with valid CSV files only, according to the format rules discussed in Comma-Separated Values (CSV).

    (,|\r?\n|^)([^",\r\n]+|"(?:[^"]|"")*")?
    Regex options: None
    Regex flavors: .NET, Java, JavaScript, PCRE, Perl, Python, Ruby

    Here is the same regular expression again in free-spacing mode:

    ( , | \r?\n | ^ )   # Capture the leading field delimiter to backref 1
(                   # Capture a single field to backref 2:
  [^",\r\n]+        #   Unquoted field
|                   #  Or:
  " (?:[^"]|"")* "  #   Quoted field (may contain escaped double quotes)
)?                  # The group is optional because fields may be empty
    Regex options: Free-spacing
    Regex flavors: .NET, Java, XRegExp, PCRE, Perl, Python, Ruby

    Using this regex and the code in Recipe 3.11, you can iterate over your CSV file and check the value of backreference 1 after each match. The necessary replacement string for each match depends on the value of this backreference. If it’s a comma, replace it with a tab character. If the backreference is empty or contains a line break, leave the value in place (i.e., do nothing, or put it back as part of a replacement string). Since CSV fields are captured to backreference 2 as part of each match, you’ll also have to put that back as part of each replacement string. The only things you’re actually replacing are the commas that are captured to backreference 1.

    Example web page with JavaScript

    The following code is a complete web page that includes two multiline text input fields, with a button labeled Replace between them. Clicking the button takes whatever string you put into the first text box (labeled Input), converts any comma delimiters to tabs with the help of the regular expression just shown, then puts the new string into the second text box (labeled Output). If you use valid CSV content as your input, it should show up in the second text box with all comma delimiters replaced with tabs. To test it, save this code into a file with the .html extension and open it in your favorite web browser:

    <html>
<head>
<title>Change CSV delimiters from commas to tabs</title>
</head>

<body>
<p>Input:</p>
<textarea id="input" rows="5" cols="75"></textarea>

<p><input type="button" value="Replace" onclick="commasToTabs()"></p>

<p>Output:</p>
<textarea id="output" rows="5" cols="75"></textarea>

<script>
function commasToTabs() {
    var input  = document.getElementById("input"),
        output = document.getElementById("output"),
        regex  = /(,|\r?\n|^)([^",\r\n]+|"(?:[^"]|"")*")?/g,
        result = "",
        match;

    while (match = regex.exec(input.value)) {
        // Check the value of backreference 1
        if (match[1] == ",") {
            // Add a tab (in place of the matched comma) and backreference
            // 2 to the result. If backreference 2 is undefined (because
            // the optional, second capturing group did not participate in
            // the match), use an empty string instead.
            result += "\t" + (match[2] || "");
        } else {
            // Add the entire match to the result
            result += match[0];
        }

        // If there is an empty match, prevent some browsers from getting
        // stuck in an infinite loop
        if (match.index == regex.lastIndex) {
            regex.lastIndex++;
        }
    }

    output.value = result;
}
</script>
</body>
</html>

Discussion

The approach prescribed by this recipe allows you to pass over each complete CSV field (including any embedded line breaks, escaped double quotes, and commas) one at a time. Each match then starts just before the next field delimiter.

The first capturing group in the regex, ‹(,|\r?\n|^)›, matches a comma, line break, or the position at the beginning of the subject string. Since the regex engine will attempt alternatives from left to right, these options are listed in the order in which they will most frequently occur in the average CSV file. This capturing group is the only part of the regex that is required to match. Therefore, it’s possible for the complete regex to match an empty string since the ‹^› anchor can match once. The value matched by this first capturing group must be checked in the code outside of the regex that replaces commas with your substitute delimiters (in this case, tabs).

We haven’t yet gotten through the entire regex, but the approach described so far is already somewhat convoluted. You might be wondering why the regex is not written to match only the commas that should be replaced with tabs. If you could do that, a simple substitution of all matched text would avoid the need for code outside of the regex to check whether capturing group 1 matched a comma or some other string. After all, it should be possible to use lookahead and lookbehind to determine whether a comma is inside or outside a quoted CSV field, right?

Unfortunately, in order for such an approach to accurately determine which commas are outside of double-quoted fields, you’d need infinite-length lookbehind, which is available in the .NET regex flavor only (see Different levels of lookbehind for a discussion of the varying lookbehind limitations). Even .NET developers should avoid a lookaround-based approach since it would add significant complexity and also make the regex slower.

Getting back to how the regex works, most of the pattern appears within the next set of parentheses: capturing group 2. This second group matches a single CSV field, including any surrounding double quotes. Unlike the previous capturing group, this one is optional in order to allow matching empty fields.

Note that group 2 within the regex contains two alternative patterns separated by the ‹|› metacharacter. The first alternative, ‹[^",\r\n]+›, is a negated character class followed by a one-or-more quantifier (‹+›) that, together, match an entire unquoted field. For this to match, the field cannot contain any double quotes, commas, or line breaks.

The second alternative within group 2, ‹"(?:[^"]|"")*"›, matches a field surrounded by double quotes. More precisely, it matches a double quote character, followed by zero or more non-double-quote characters or repeated (escaped) double quotes, followed by a closing double quote. The ‹*› quantifier at the end of the noncapturing group continues repeating the two options within the group until it reaches a double quote that is not repeated and therefore ends the field.

Assuming you’re working with a valid CSV file, the first match found by this regex should occur at the beginning of the subject string, and each subsequent match should occur immediately after the end of the last match.

See Also

Recipe 9.12 describes how to reuse the regex in this recipe to extract CSV fields from a specific column.

Techniques used in the regular expressions in this recipe are discussed in Chapter 2. Recipe 2.2 explains how to match nonprinting characters. Recipe 2.3 explains character classes. Recipe 2.5 explains anchors. Recipe 2.8 explains alternation. Recipe 2.9 explains grouping. Recipe 2.12 explains repetition.