Bash嵌套循环遍历目录执行命令：最后目录重复循环4次

Hey there! Let's break down your problem into two solvable parts: extracting the unique segment from your fastq filenames with sed, and fixing that annoying bug where the last directory gets processed 4 times.

1. Extracting the Unique Filename Segment #

Your filenames follow the pattern name_S#_L001_R1_001.fastq.gz—the unique part is everything before _L001. Here's a sed command that snips off the redundant suffix cleanly:

# Example usage with a sample filename
echo "sampleA_S3_L001_R1_001.fastq.gz" | sed -E 's/(_L001_R[12]_001\.fastq\.gz)//'

What this does: #

• -E enables extended regular expressions for cleaner syntax.
• s/(_L001_R[12]_001\.fastq\.gz)// matches the fixed suffix (accounting for both R1 and R2 reads) and replaces it with nothing, leaving only your unique name_S# segment.

You can embed this directly into your loop to capture the unique name for each file.

2. Fixing the Repeated Last Directory Loop #

The root cause of your last directory repeating 4 times is almost certainly one of two issues:

• Variable name collision: Using the same variable (like f) for both your outer directory loop and inner file loop, overwriting the outer loop's value.
• Not returning to the parent directory: After cding into a subdirectory, you don't navigate back, which can mess with subsequent loop iterations.

Here are two robust fixes:

Fix Option 1: Use Distinct Variables & Reset Working Directory #

This approach saves your original directory, uses unique variable names, and ensures you return to the parent after processing each subdirectory:

# Save your starting directory to return to later
original_dir=$(pwd)

# Outer loop: iterate over all items in your target path, filter for directories
for dir in /completepath/*
do
    # Only process if the item is a directory
    if [ -d "$dir" ]; then
        # Enter the directory (skip if cd fails)
        cd "$dir" || continue
        
        # Inner loop: process each fastq.gz file
        for file in *.fastq.gz
        do
            # Extract the unique filename segment
            unique_name=$(echo "$file" | sed -E 's/(_L001_R[12]_001\.fastq\.gz)//')
            
            # Add your processing command here (e.g., alignment, renaming)
            echo "Processing $unique_name in directory: $(basename "$dir")"
        done
        
        # Return to your original starting directory
        cd "$original_dir"
    fi
done

Fix Option 2: Use a Subshell to Isolate Directory Changes #

A cleaner alternative is to wrap the directory-specific code in a subshell (()). This way, cd only affects the subshell, and your main shell stays in the original directory—no need to manually navigate back:

for dir in /completepath/*
do
    if [ -d "$dir" ]; then
        # Subshell: all commands here run in a separate shell context
        (
            cd "$dir" || exit  # Exit subshell if cd fails
            for file in *.fastq.gz
            do
                unique_name=$(echo "$file" | sed -E 's/(_L001_R[12]_001\.fastq\.gz)//')
                echo "Processing $unique_name in directory: $(basename "$dir")"
                # Add your processing commands here
            done
        )
        # Subshell exits, main shell is still in original directory
    fi
done

Why this fixes the repetition: #

By using unique variable names (dir for directories, file for files) we avoid overwriting the outer loop's iterator. The subshell method adds extra isolation, so directory changes never affect the main loop's state.

内容的提问来源于stack exchange，提问作者K-MO