When files from differnt processess have the same name, workflow outputs only copy one file

[GabeGEL]

Bug report

Expected behavior and actual behavior

I have two processes that both create files with the same name.
I expect to be able to publish both files and update the names in the output block.

Instead, only the second file gets output. eg:

output {
    profiles {
        mode 'copy'
        path { out ->
            // A file named "pyspy.prof" from the PROFILE_A process
            out.profile_a >> "pyspy/${out.meta.id}.profile_a.pyspy.prof"
            // A file named "pyspy.prof" from the PROFILE_B process
            out.profile_b >> "pyspy/${out.meta.id}.profile_b.pyspy.prof"
}}}

I would expect to get outputs such as:

results
└── pyspy
    ├── case001.profile_a.pyspy.prof
    ├── case001.profile_b.pyspy.prof
    ├── case002.profile_a.pyspy.prof
    └── case002.profile_b.pyspy.prof

But what I actually get is:

results
└── pyspy
    ├── case001.profile_b.pyspy.prof
    └── case002.profile_b.pyspy.prof

Furthermore, the file content is incorrect:

head ./results/pyspy/*

==> ./results/pyspy/case001.profile_b.pyspy.prof <==
PROFILE_A

==> ./results/pyspy/case002.profile_b.pyspy.prof <==
PROFILE_B

I can confirm that the files were not mis-attributed in the join. The following is the output from the toy example below:

[meta:[id:case001], profile_a:/Users/gabrielaldam/Projects/consolidate_abtiering/fail/work/87/0de5c7ec9482e8f25a0f8d0e553bc6/pyspy.prof, profile_b:/Users/gabrielaldam/Projects/consolidate_abtiering/fail/work/89/ad6c7974848a11b1fba224b09f0023/pyspy.prof]
[meta:[id:case002], profile_a:/Users/gabrielaldam/Projects/consolidate_abtiering/fail/work/56/5827cde5334bde58db0053e8300bc8/pyspy.prof, profile_b:/Users/gabrielaldam/Projects/consolidate_abtiering/fail/work/df/012645b58f7fe7a0b7ca189b6f02ed/pyspy.prof]

And the file content:

cat /Users/gabrielaldam/Projects/consolidate_abtiering/fail/work/87/0de5c7ec9482e8f25a0f8d0e553bc6/pyspy.prof
PROFILE_A
cat /Users/gabrielaldam/Projects/consolidate_abtiering/fail/work/89/ad6c7974848a11b1fba224b09f0023/pyspy.prof
PROFILE_B
cat /Users/gabrielaldam/Projects/consolidate_abtiering/fail/work/56/5827cde5334bde58db0053e8300bc8/pyspy.prof
PROFILE_A
cat /Users/gabrielaldam/Projects/consolidate_abtiering/fail/work/df/012645b58f7fe7a0b7ca189b6f02ed/pyspy.prof
PROFILE_B

Steps to reproduce the problem

The following example can reproduce the issue:

workflow {
    main:
    ch_cases = channel.of('case001','case002')
    PROFILE_A(ch_cases)
    PROFILE_B(ch_cases)

    ch_profiles = PROFILE_A.out.profiles.join(PROFILE_B.out.profiles)
        .map { caseid, profile_a, profile_b ->
            [
                meta:[ id: caseid ],
                profile_a: profile_a,
                profile_b: profile_b
            ]
        }
    ch_profiles.view()


    publish:
    profiles = ch_profiles
}

output {
    profiles {
        mode 'copy'
        path { out ->
            out.profile_a >> "pyspy/${out.meta.id}.profile_a.pyspy.prof"
            out.profile_b >> "pyspy/${out.meta.id}.profile_b.pyspy.prof"
        }
    }
}

process PROFILE_A {
    input:
    val caseid

    script:
    """
    echo "PROFILE_A" > pyspy.prof
    """

    output:
    tuple val(caseid), path("pyspy.prof"), emit: profiles
}

process PROFILE_B {
    input:
    val caseid

    script:
    """
    echo "PROFILE_B" > pyspy.prof
    """

    output:
    tuple val(caseid), path("pyspy.prof"), emit: profiles
}

Program output

Nov-27 15:44:12.308 [main] DEBUG nextflow.cli.Launcher - $> nextflow run ./main.nf
Nov-27 15:44:12.356 [main] DEBUG nextflow.cli.CmdRun - N E X T F L O W  ~  version 25.10.0
Nov-27 15:44:12.410 [main] DEBUG nextflow.plugin.PluginsFacade - Setting up plugin manager > mode=prod; embedded=false; plugins-dir=/Users/gabrielaldam/.nextflow/plugins; core-plugins: nf-amazon@3.4.1,nf-azure@1.20.2,nf-cloudcache@0.5.0,nf-codecommit@0.5.0,nf-console@1.3.0,nf-google@1.23.3,nf-k8s@1.2.2,nf-tower@1.17.1,nf-wave@1.16.1
Nov-27 15:44:12.425 [main] INFO  o.pf4j.DefaultPluginStatusProvider - Enabled plugins: []
Nov-27 15:44:12.426 [main] INFO  o.pf4j.DefaultPluginStatusProvider - Disabled plugins: []
Nov-27 15:44:12.427 [main] INFO  org.pf4j.DefaultPluginManager - PF4J version 3.12.0 in 'deployment' mode
Nov-27 15:44:12.436 [main] DEBUG nextflow.util.RetryConfig - Missing nextflow session - using default retry config
Nov-27 15:44:12.534 [main] INFO  org.pf4j.AbstractPluginManager - No plugins
Nov-27 15:44:12.535 [main] DEBUG nextflow.plugin.PluginsFacade - Plugins default=[]
Nov-27 15:44:12.535 [main] DEBUG nextflow.plugin.PluginsFacade - Plugins resolved requirement=[]
Nov-27 15:44:12.545 [main] DEBUG n.secret.LocalSecretsProvider - Secrets store: /Users/gabrielaldam/.nextflow/secrets/store.json
Nov-27 15:44:12.550 [main] DEBUG nextflow.secret.SecretsLoader - Discovered secrets providers: [nextflow.secret.LocalSecretsProvider@67c277a0] - activable => nextflow.secret.LocalSecretsProvider@67c277a0
Nov-27 15:44:12.572 [main] DEBUG nextflow.cli.CmdRun - Applied DSL=2 by global default
Nov-27 15:44:12.588 [main] DEBUG nextflow.cli.CmdRun - Launching `./main.nf` [festering_mahavira] DSL2 - revision: d5c3127ffd
Nov-27 15:44:12.619 [main] DEBUG nextflow.Session - Session UUID: d0d03db4-48d4-48a0-8f82-5cc3c25072ee
Nov-27 15:44:12.619 [main] DEBUG nextflow.Session - Run name: festering_mahavira
Nov-27 15:44:12.619 [main] DEBUG nextflow.Session - Executor pool size: 8
Nov-27 15:44:12.624 [main] DEBUG nextflow.file.FilePorter - File porter settings maxRetries=3; maxTransfers=50; pollTimeout=null
Nov-27 15:44:12.627 [main] DEBUG nextflow.util.ThreadPoolBuilder - Creating thread pool 'FileTransfer' minSize=10; maxSize=24; workQueue=LinkedBlockingQueue[-1]; allowCoreThreadTimeout=false
Nov-27 15:44:12.648 [main] DEBUG nextflow.cli.CmdRun -
  Version: 25.10.0 build 10289
  Created: 22-10-2025 16:26 UTC (17:26 BST)
  System: Mac OS X 26.1
  Runtime: Groovy 4.0.28 on OpenJDK 64-Bit Server VM 17.0.15+6-LTS
  Encoding: UTF-8 (UTF-8)
  Process: 81021@GEL-G9C747KD9H [10.255.51.115]
  CPUs: 8 - Mem: 16 GB (76 MB) - Swap: 6 GB (1.1 GB)
Nov-27 15:44:12.657 [main] DEBUG nextflow.Session - Work-dir: /Users/gabrielaldam/Projects/consolidate_abtiering/fail/work [Mac OS X]
Nov-27 15:44:12.657 [main] DEBUG nextflow.Session - Script base path does not exist or is not a directory: /Users/gabrielaldam/Projects/consolidate_abtiering/fail/bin
Nov-27 15:44:12.662 [main] DEBUG nextflow.executor.ExecutorFactory - Extension executors providers=[]
Nov-27 15:44:12.666 [main] DEBUG nextflow.Session - Observer factory (v2): LinObserverFactory
Nov-27 15:44:12.667 [main] DEBUG nextflow.Session - Observer factory (v2): DefaultObserverFactory
Nov-27 15:44:12.686 [main] DEBUG nextflow.cache.CacheFactory - Using Nextflow cache factory: nextflow.cache.DefaultCacheFactory
Nov-27 15:44:12.691 [main] DEBUG nextflow.util.CustomThreadPool - Creating default thread pool > poolSize: 9; maxThreads: 1000
Nov-27 15:44:12.721 [main] DEBUG nextflow.Session - Session start
Nov-27 15:44:13.042 [main] DEBUG nextflow.script.ScriptRunner - > Launching execution
Nov-27 15:44:13.104 [main] DEBUG nextflow.executor.ExecutorFactory - << taskConfig executor: null
Nov-27 15:44:13.105 [main] DEBUG nextflow.executor.ExecutorFactory - >> processorType: 'local'
Nov-27 15:44:13.109 [main] DEBUG nextflow.executor.Executor - [warm up] executor > local
Nov-27 15:44:13.112 [main] DEBUG n.processor.LocalPollingMonitor - Creating local task monitor for executor 'local' > cpus=8; memory=16 GB; capacity=8; pollInterval=100ms; dumpInterval=5m
Nov-27 15:44:13.113 [main] DEBUG n.processor.TaskPollingMonitor - >>> barrier register (monitor: local)
Nov-27 15:44:13.128 [main] DEBUG nextflow.processor.TaskProcessor - Creating process 'PROFILE_A': maxForks=0; fair=false; array=0
Nov-27 15:44:13.148 [main] DEBUG nextflow.executor.ExecutorFactory - << taskConfig executor: null
Nov-27 15:44:13.148 [main] DEBUG nextflow.executor.ExecutorFactory - >> processorType: 'local'
Nov-27 15:44:13.149 [main] DEBUG nextflow.processor.TaskProcessor - Creating process 'PROFILE_B': maxForks=0; fair=false; array=0
Nov-27 15:44:13.169 [main] DEBUG nextflow.Session - Workflow process names [dsl2]: PROFILE_A, PROFILE_B
Nov-27 15:44:13.170 [main] DEBUG nextflow.Session - Igniting dataflow network (3)
Nov-27 15:44:13.170 [main] DEBUG nextflow.processor.TaskProcessor - Starting process > PROFILE_A
Nov-27 15:44:13.170 [main] DEBUG nextflow.processor.TaskProcessor - Starting process > PROFILE_B
Nov-27 15:44:13.171 [main] DEBUG nextflow.script.ScriptRunner - Parsed script files:
  Script_50d6b2ddf702c0dd: /Users/gabrielaldam/Projects/consolidate_abtiering/fail/main.nf
Nov-27 15:44:13.171 [main] DEBUG nextflow.script.ScriptRunner - > Awaiting termination
Nov-27 15:44:13.171 [main] DEBUG nextflow.Session - Session await
Nov-27 15:44:13.246 [Task submitter] DEBUG n.executor.local.LocalTaskHandler - Launch cmd line: /bin/bash -ue .command.run
Nov-27 15:44:13.247 [Task submitter] INFO  nextflow.Session - [87/0de5c7] Submitted process > PROFILE_A (1)
Nov-27 15:44:13.252 [Task submitter] DEBUG n.executor.local.LocalTaskHandler - Launch cmd line: /bin/bash -ue .command.run
Nov-27 15:44:13.252 [Task submitter] INFO  nextflow.Session - [df/012645] Submitted process > PROFILE_B (2)
Nov-27 15:44:13.256 [Task submitter] DEBUG n.executor.local.LocalTaskHandler - Launch cmd line: /bin/bash -ue .command.run
Nov-27 15:44:13.256 [Task submitter] INFO  nextflow.Session - [89/ad6c79] Submitted process > PROFILE_B (1)
Nov-27 15:44:13.260 [Task submitter] DEBUG n.executor.local.LocalTaskHandler - Launch cmd line: /bin/bash -ue .command.run
Nov-27 15:44:13.260 [Task submitter] INFO  nextflow.Session - [56/5827cd] Submitted process > PROFILE_A (2)
Nov-27 15:44:13.323 [Task monitor] DEBUG n.processor.TaskPollingMonitor - Task completed > TaskHandler[id: 1; name: PROFILE_A (1); status: COMPLETED; exit: 0; error: -; workDir: /Users/gabrielaldam/Projects/consolidate_abtiering/fail/work/87/0de5c7ec9482e8f25a0f8d0e553bc6]
Nov-27 15:44:13.324 [Task monitor] DEBUG nextflow.util.ThreadPoolBuilder - Creating thread pool 'TaskFinalizer' minSize=10; maxSize=24; workQueue=LinkedBlockingQueue[-1]; allowCoreThreadTimeout=false
Nov-27 15:44:13.325 [Task monitor] DEBUG n.processor.TaskPollingMonitor - Task completed > TaskHandler[id: 3; name: PROFILE_B (2); status: COMPLETED; exit: 0; error: -; workDir: /Users/gabrielaldam/Projects/consolidate_abtiering/fail/work/df/012645b58f7fe7a0b7ca189b6f02ed]
Nov-27 15:44:13.326 [Task monitor] DEBUG n.processor.TaskPollingMonitor - Task completed > TaskHandler[id: 2; name: PROFILE_B (1); status: COMPLETED; exit: 0; error: -; workDir: /Users/gabrielaldam/Projects/consolidate_abtiering/fail/work/89/ad6c7974848a11b1fba224b09f0023]
Nov-27 15:44:13.326 [Task monitor] DEBUG n.processor.TaskPollingMonitor - Task completed > TaskHandler[id: 4; name: PROFILE_A (2); status: COMPLETED; exit: 0; error: -; workDir: /Users/gabrielaldam/Projects/consolidate_abtiering/fail/work/56/5827cde5334bde58db0053e8300bc8]
Nov-27 15:44:13.360 [main] DEBUG nextflow.Session - Session await > all processes finished
Nov-27 15:44:13.373 [Actor Thread 6] DEBUG nextflow.util.ThreadPoolBuilder - Creating thread pool 'PublishDir' minSize=10; maxSize=24; workQueue=LinkedBlockingQueue[-1]; allowCoreThreadTimeout=false
Nov-27 15:44:13.419 [Task monitor] DEBUG n.processor.TaskPollingMonitor - <<< barrier arrives (monitor: local) - terminating tasks monitor poll loop
Nov-27 15:44:13.419 [main] DEBUG nextflow.Session - Session await > all barriers passed
Nov-27 15:44:13.421 [main] DEBUG nextflow.util.ThreadPoolManager - Thread pool 'TaskFinalizer' shutdown completed (hard=false)
Nov-27 15:44:13.421 [main] DEBUG nextflow.util.ThreadPoolManager - Thread pool 'PublishDir' shutdown completed (hard=false)
Nov-27 15:44:13.424 [main] DEBUG n.trace.WorkflowStatsObserver - Workflow completed > WorkflowStats[succeededCount=4; failedCount=0; ignoredCount=0; cachedCount=0; pendingCount=0; submittedCount=0; runningCount=0; retriesCount=0; abortedCount=0; succeedDuration=6ms; failedDuration=0ms; cachedDuration=0ms;loadCpus=0; loadMemory=0; peakRunning=4; peakCpus=4; peakMemory=0; ]
Nov-27 15:44:13.627 [main] DEBUG nextflow.cache.CacheDB - Closing CacheDB done
Nov-27 15:44:13.635 [main] DEBUG nextflow.util.ThreadPoolManager - Thread pool 'FileTransfer' shutdown completed (hard=false)
Nov-27 15:44:13.635 [main] DEBUG nextflow.script.ScriptRunner - > Execution complete -- Goodbye

Environment

• Nextflow version: nextflow version 25.10.0.10289
• Java version: OpenJDK 64-Bit Server VM Homebrew (build 22, mixed mode, sharing)
• Operating system: macOS (26.1 (25B78))
• Bash version: zsh 5.9 (arm64-apple-darwin25.0)

Additional context

(Add any other context about the problem here)

[bentsherman]

Thanks for reporting, and for the clean reprex

The output path closure creates a mapping of relative filename -> target path, which does not encode anything about specific task directories. So if you publish two files with the same relative filename but from different tasks, the mapping will only remember the second publish statement

This is partly a consequence of reusing the existing publishDir logic -- the output path closure basically uses the publishDir saveAs option under the hood which uses the relative filename. #6561 is another downside of re-using PublishDir. So we might need to just re-implement the publishing logic in a new class that is designed from the ground-up for workflow outputs.

The mapping is evaluated for every channel value, so it should just encode the full source path rather than the relative filename.