使用CDO与Shell脚本将小时级NetCDF文件合并为日级文件：构造合并命令的实现方法

First off, let's fix a common pitfall with your initial variable setup: storing ls output in a regular variable can cause issues if your filenames contain spaces or special characters. Instead, use a shell array to safely hold all your NetCDF files—this is way more reliable:

# Use a shell array to store all .nc files (sorted alphabetically by default)
files=(*.nc)

Now, to merge every 24 hourly files into a single daily file, we can use a for loop with a step size of 24. Here's a complete, robust implementation:

# Check if we have any files to process
if [ ${#files[@]} -eq 0 ]; then
    echo "No .nc files found in the current directory!"
    exit 1
fi

# Loop through the array in batches of 24
for ((i=0; i<${#files[@]}; i+=24)); do
    # Calculate the end index of the current batch (avoid exceeding array length)
    end=$((i+23))
    if [ $end -ge ${#files[@]} ]; then
        end=$((${#files[@]}-1))
        echo "Warning: Final batch has fewer than 24 files (from index $i to $end)"
    fi

    # Extract the current batch of files
    batch_files=("${files[@]:i:24}")

    # Define a clean, sorted output filename (customize this to your needs)
    # Example: day_01.nc, day_02.nc, etc.
    output_file="day_$(printf "%02d" $((i/24 +1))).nc"

    # Construct and run the CDO merge command
    echo "Merging files ${batch_files[*]} into $output_file..."
    cdo merge "${batch_files[@]}" "$output_file"
done

Let's break this down: #

• Array Setup: files=(*.nc) creates an array where each element is a separate filename, completely avoiding issues with spaces or special characters in filenames.
• Batch Loop: The for ((i=0; i<${#files[@]}; i+=24)) loop increments by 24 each iteration, so we process one full day's worth of hourly files per run.
• Edge Case Handling: The code checks if the final batch has fewer than 24 files and warns you, instead of failing silently.
• Output Naming: Using printf "%02d" ensures day numbers are zero-padded (e.g., 01 instead of 1) for consistent filesystem sorting.
• CDO Command: "${batch_files[@]}" expands to each filename as a separate argument—exactly what CDO expects for its merge operation.

If you really want to stick with your original variable (not recommended): #

If you insist on using files=$(ls *.nc) (note: this is not ideal for filenames with spaces), you can convert it to an array first:

files_str=$(ls *.nc)
read -ra files <<< "$files_str"
# Then use the same loop as above

Just remember: using ls in scripts is generally discouraged because it can break with non-standard filenames. The array approach with *.nc is the safer, more robust choice.

内容的提问来源于stack exchange，提问作者João Janeiro