[
https://issues.apache.org/jira/browse/LANG-502?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12752010#action_12752010
]
Henri Yandell commented on LANG-502:
------------------------------------
Given that we have a fast replace method; one question would be whether to add
a private underlying method and a last argument of boolean ignoreCase. Then
have it make a couple of string variable copies and if true make the copies
lowercase. A very slight performance hit for the current replace (two variable
declarations to existing objects and an if statement on a boolean) and not much
increase in terms of maintenance.
Other option would be to decide this isn't a common enough need for performance
to be critical and using regex with /i is the way to go.
> new method StringUtils.replaceIgnoreCase (with patch)
> -----------------------------------------------------
>
> Key: LANG-502
> URL: https://issues.apache.org/jira/browse/LANG-502
> Project: Commons Lang
> Issue Type: Improvement
> Environment: all
> Reporter: Flo
> Priority: Minor
> Fix For: 3.0
>
>
> Method implementation:
> /**
> * Searches for all appearances of <code>searchString</code> (ignoring
> case) in <code>text</code> and replaces them by <code>replacement</code>.
> * The difference to {...@link String#replace(CharSequence,
> CharSequence)} and {...@link StringUtils#replace(String, String, String)} is
> that this implementation ignores case.
> *
> * @param text The text in which to do replacements.
> * @param searchString The string to remove from the text (ignoring
> case).
> * @param replacement The string to put instead of the searchString.
> * @return A new string with all searchString replaced by replacement.
> * @author frickert
> */
> public static String replaceIgnoreCase(String text, String
> searchString, String replacement)
> {
> String lowerCaseText = text.toLowerCase();
> String lowerCaseSearchString = searchString.toLowerCase();
> StringBuilder sb = new StringBuilder(text);
> int searchStart = 0;
> final int modifierPerReplacement = replacement.length() -
> searchString.length();
> int sbDrift = 0; // by doing replacements in sb, sb and the
> text drift off in length and index in case the searchString and the
> replacement are of different length
> int finding = lowerCaseText.indexOf(lowerCaseSearchString,
> searchStart);
> while (finding >= 0)
> {
> sb.replace(finding + sbDrift, finding + sbDrift +
> searchString.length(), replacement);
> sbDrift += modifierPerReplacement;
> searchStart = finding + searchString.length();
> finding = lowerCaseText.indexOf(lowerCaseSearchString,
> searchStart);
> }
> return sb.toString();
> }
> test cases:
> public void testReplaceIgnoreCase() throws Throwable {
> String is;
> is = CommonHelpers.replaceIgnoreCase("bobOBOBobOB", "Bob", "Flo");
> assertEquals("search really ignores case", "FloOFlooFlo", is);
> is = CommonHelpers.replaceIgnoreCase("bobOBOBobOB", "Bob", "Flo");
> assertEquals("replacement does care about case", "FloOFlooFlo", is);
> is = CommonHelpers.replaceIgnoreCase("bob bob bob", "Bob", "Florian");
> assertEquals("length difference of searchString and replacement > 0",
> "Florian Florian Florian", is);
> is = CommonHelpers.replaceIgnoreCase("bob bob bob", "Bob", "Ed");
> assertEquals("length difference of searchString and replacement < 0",
> "Ed Ed Ed", is);
> is = CommonHelpers.replaceIgnoreCase("GROSS und klein", "und", "&");
> assertEquals("originals case is preserved in not replace chars",
> "GROSS & klein", is);
> }
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
