Search through .xml files for text that exists anywhere in the file

• RootDir\172\00ad8dc-bff\25.0.xml

• RootDir- The root directory for all my .xml files.

• 172- The sub directory of the "RootDir" directory. There are many such directories (I gave 172 as an example)

• 00ad8dc-bff- The sub directory under "172" directory. There are many such directories.

• 25.0.xml- At this level I have my target .xml files in which the search should be performed.

• There is a folder with name "processes".

• The file name is "draft.xml".

• Check if given path A is a directory.

• If path A is a directory, we attempt to get an array of paths A/B* inside A.

• For each of the path A/B in A/B, we attempt to get an array of paths A/B/C inside each A/B.

• For each of the path A/B/C in A/B/C*, we check if that is a directory.

• If A/B/C is a directory, we attempt to get an array of paths A/B/C/D* and then we determine the latest modified file A/B/C/D. If there are multiple latest files, then by default the first iterated is our 'target file'.

• If A/B/C/D does not contain processes or draft in its name, then we create a BufferedReader for it, referenced by a static variable, and then proceed with search().

• Search results are saved in a static variable uuids.

• If A/B/C/D contains processes or draft in its name, we skip processing in the current A/B/C directory and go to the next one.

• Nested indentation runs deep, because every step is nested in the preceding if or for statement.

• Even if you want to clump the logic together, you should try to check for terminal conditions first. For example, If step 1 above is not true, then your program can pretty much exit. Steps 2 to 7 need not be nested inside the if statement in step 1.

• Also, you may have only up to two levels of directories, but what happens when you have more in the future?

• static variables are used over properly designed method signatures, which may will get confusing once you have more 'arguments' to deal with throughout your code.

• Do you really want to ignore \$n\$ files when \$n + 1\$ have the same latest modified timestamps in step 5?

• I am not sure if your logic for step 7 is right: We think we found the latest file, but if the file name is not the one we want, we go to the next sibling directory (i.e. A/B/C1/ after A/B/C2/)? Do we even want to check that maybe the second latest file (see above issue on point 5) should be searched as well?

public static Set getFileCandidates(final String rootDirectory);

// assuming the seach term is fixed
public static List getMatchingLines(final File file);

public static void main(String[] args) {
    String rootDirectory = args[0]; // for example
    final List results = new ArrayList<>();
    for (final File fileCandidate : getFileCandidates(rootDirectory)) {
        results.addAll(getMatchingLines(fileCandidate));
    }
    System.out.println(results);
}

• If current line is the first to contain "uuid", 'save' this line (tempUUIDLine) and continue to process other lines not containing "uuid".

• If the current line contains searchQuery, print tempUUIDLine.

• Your code is broken, even though you may think it is working to the best of your knowledge. If this is a case, do consider posting a SO question to fix how your search() is being done first, before posting a follow-up review here.

• If you end up having to scan through multiple lines forwards/backwards within the file to do your data extraction, you really might be better off parsing the XML file into a Document after all to get your results using a more standardized solution.

!lastModifiedFile.getName().contains("processes") && 
    !lastModifiedFile.getName().contains("draft")