List particular file types from dynamic folders in java

[velan krishnan]

Hi
I have folder structure like rootfolder/dynamic folders/test

test folder has XML files and text files.
I need to to list only XML files.
Need Something like below
rootfolder/**/test*.xml

[fred rosenberger]

What have you tried?  What did it do different than what you expected?

[Tim Holloway]

Welcome the the Ranch, Velan!

I'm not sure that you want  to "list" files so much as enumerate them or collect their names/paths. To "list" files in Linux, a command line would suffice:

find rootfolder -name "test*.xml"

Traditionally, to enumerate and/or collect in Java, you'd use the file list() method with a filter that gives a yes/no indication for each file scanned to indicate whether it matched your selected pattern.

Alternatively these days, there are evil things you can do with Lamba expressions to the same effect, but I'll leave the details to others.

[Ron McLeod]

Take a look at Files#find.  Here's an example of what I had done before:

import java.io.IOException;
import java.nio.file.FileVisitOption;
import java.nio.file.Files;
import java.nio.file.Paths;
import java.util.regex.Pattern;
import java.util.stream.Collectors;
import org.junit.jupiter.api.Test;

public class RonTest {
    @Test
    void fileFinderTest() throws IOException {
        final var rootPath = Paths.get("/home/mcleod/.m2/repository");
        final var matcher = Pattern.compile("smallrye-common-annotation.*\\.pom").asMatchPredicate();
        final var pathDepth = 10;

        var files = Files.find(rootPath,
                pathDepth,
                (path, attr) -> attr.isRegularFile()
                        && matcher.test(path.getName(path.getNameCount() - 1).toString()),
                FileVisitOption.FOLLOW_LINKS)
                .collect(Collectors.toList());

        files.forEach(System.out::println);
    }
}

/home/mcleod/.m2/repository/io/smallrye/.../smallrye-common-annotation-1.6.0.pom /home/mcleod/.m2/repository/io/smallrye/.../smallrye-common-annotation-1.13.1.pom /home/mcleod/.m2/repository/io/smallrye/.../smallrye-common-annotation-1.10.0.pom /home/mcleod/.m2/repository/io/smallrye/.../smallrye-common-annotation-1.0.2.pom /home/mcleod/.m2/repository/io/smallrye/.../smallrye-common-annotation-1.7.1.pom /home/mcleod/.m2/repository/io/smallrye/.../smallrye-common-annotation-1.8.0.pomI shortend the paths with ...

[Carey Brown]

So for Velan it might be something like this:

        var files = Files.find(rootPath,
                pathDepth,
                (path, attr) -> attr.isRegularFile()
                        && path.getName(path.getNameCount() - 2).toString().equals("test")
                        && path.getName(path.getNameCount() - 1).toString().endsWith(".xml"),
                FileVisitOption.FOLLOW_LINKS)
                .collect(Collectors.toList());

[Tim Holloway]

Carey Brown wrote:So for Velan it might be something like this:

        var files = Files.find(rootPath,
                pathDepth,
                (path, attr) -> attr.isRegularFile()
                        && path.getName(path.getNameCount() - 2).toString().equals("test")
                        && path.getName(path.getNameCount() - 1).toString().endsWith(".xml"),
                FileVisitOption.FOLLOW_LINKS)
                .collect(Collectors.toList());

I think that the cleaner solution (assuming that regexes are ever "clean"!) is more like this:

Pattern filePattern = Pattern.compile(".*/test.*\\.xml$");
....
               (path, attr) -> attr.isRegularFile()
                       &&filePattern.match(path), …

That's the concept at least. Any resemblance between that and actual working Java regex code can be considered entirely accidental.

[Carey Brown]

Tim, the OP made a typo in his problem description, "test" is a directory and not part of a file name.

[Tim Holloway]

Carey Brown wrote:Tim, the OP made a typo in his problem description, "test" is a directory and not part of a file name.

Ah well, just changes the regex a little.

[Carey Brown]

Tim Holloway wrote:

Carey Brown wrote:Tim, the OP made a typo in his problem description, "test" is a directory and not part of a file name.

Ah well, just changes the regex a little.

Wouldn't that also mean getting a different segment of the path?

[Tim Holloway]

Then you'd just regex against the full path and not bother to rip out the filename part.

But I don't assume that's necessary, since upon more closely reviewing the problem description, it actually reads more like files in the form:

rootfolder/dynamic folders/test/test001.xml
rootfolder/dynamic folders/test/test002.xml
rootfolder/dynamic folders/test/test003.xml
rootfolder/dynamic folders/test/testabcd.xml

etc.

[Ron McLeod]

If you filter based on the string representation of the full path, then you will need to consider the path separator character in your pattern (/ for Linux, \ for Windows).

jshell> Files.find(Path.of("D:\\ASN.1\\OSS"),        2,        (path, attr) -> attr.isRegularFile(),        FileVisitOption.FOLLOW_LINKS)    .forEach(System.out::println); D:\ASN.1\OSS\asn1pjav_win64_trial_v8.exe D:\ASN.1\OSS\asn1step_win64_trial_v10.exe D:\ASN.1\OSS\ossinfo1steptrial D:\ASN.1\OSS\ossinfojavatrial

[Tim Holloway]

Then you're using the wrong Path method. Java supports a Universal (Unix-like) path notation where the separator is always a forward ("real") slash, and never a backslash (Windows) colon (Legacy Macintosh), '>' (Primos) or possibly non-terminal dot (IBM zOS). It's specifically to assist in "write-once/run-anywhere" and it's the format that I always promote. Since I have spent literally years as a full-time developer using Windows on the Desktop to produce apps to run on Linux and Solaris and I didn't want to have to debug code that worked differently on the desktop than in production.

Although as far as that goes, if you cannot resolve to an OS-independent form, use [\\/] as the path separator in the regex.