如何让grep匹配bash脚本数组变量中的所有目标数值？

The issue in your script is that when you use grep $threshold dffile, Bash only expands $threshold to the first element of the array ("2%"), so grep never looks for "6%" which is present in your df output. Here's how to fix it and improve the script:

Why It Fails #

In Bash, referencing an array with just the variable name ($threshold) returns only the first element. To access all elements, you need to use syntax like ${threshold[@]} or ${threshold[*]}, but you also need to format them correctly for grep to interpret as multiple patterns.

Solution 1: Use Extended Regex with OR Separators #

Modify the grep line to join all array elements into a single regex pattern separated by | (OR):

grep -E "$(IFS='|'; echo "${threshold[*]}")" dffile >> thresh

• IFS='|': Temporarily sets the internal field separator to | so when we expand ${threshold[*]}, elements are joined with |.
• -E: Enables extended regex, so grep recognizes | as an OR operator.

Solution 2: Read Patterns from Process Substitution #

If you prefer to avoid regex (e.g., if your patterns might contain special regex characters), use grep's -f option to read patterns from a process substitution:

grep -f <(printf "%s\n" "${threshold[@]}") dffile >> thresh

• printf "%s\n" "${threshold[@]}": Prints each array element on its own line.
• <(...): Creates a temporary file-like object that grep reads patterns from via -f.

Corrected Full Script #

Here's your script with the first solution applied, plus optional cleanup to remove unnecessary intermediate files:

#!/bin/bash
threshold=("2%" "6%")

# Combine df output, grep matching thresholds, extract mount points, and process
df -h | grep -E "$(IFS='|'; echo "${threshold[*]}")" | awk '{print $6}' | while IFS= read -r mount_point
do
    find "$mount_point" -xdev -size +40M -exec ls -lah {} \; | head -n 10
done

This version eliminates intermediate files (dffile, thresh, finding1) by piping commands directly, making the script more efficient and cleaner.

Testing with Your df Output #

Given your test server's df output, the corrected grep will match the line with 6% /, extract / as the mount point, and run the find command on / to list large files.

内容的提问来源于stack exchange，提问作者Blake Smreker