Delete directories that start with YYYY-MM-DD with dates earlier than today

cleanup-date-directories.sh

#!/bin/bash

# Script to delete directories with date prefixes older than current date
# Usage: ./cleanup_dated_dirs.sh [--dry-run] [directory_path]

# Default values
DRY_RUN=false
TARGET_DIR="."

# Parse command line arguments
while [[ $# -gt 0 ]]; do
	case $1 in
		--dry-run)
			DRY_RUN=true
			shift
			;;
		-h|--help)
			echo "Usage: $0 [--dry-run] [directory_path]"
			echo ""
			echo "Options:"
			echo "  --dry-run     Show what would be deleted without actually deleting"
			echo "  -h, --help    Show this help message"
			echo ""
			echo "Arguments:"
			echo "  directory_path    Directory to scan (default: current directory)"
			echo ""
			echo "Example:"
			echo "  $0 --dry-run /path/to/logs"
			exit 0
			;;
		*)
			TARGET_DIR="$1"
			shift
			;;
	esac
done

# Check if target directory exists
if [[ ! -d "$TARGET_DIR" ]]; then
	echo "Error: Directory '$TARGET_DIR' does not exist"
	exit 1
fi

# Get current date in YYYY-MM-DD format
CURRENT_DATE=$(date +%Y-%m-%d)
echo "Current date: $CURRENT_DATE"
echo "Target directory: $TARGET_DIR"

if [[ "$DRY_RUN" == true ]]; then
	echo "*** DRY RUN MODE - No directories will be deleted ***"
fi

echo ""

# Counters
deleted_count=0
skipped_count=0

# Function to check if a string is a valid date in YYYY-MM-DD format
is_valid_date() {
	local date_str="$1"
	
	echo "    Validating date: '$date_str'"
	
	# Check format with regex
	if [[ ! "$date_str" =~ ^[0-9]{4}-[0-9]{2}-[0-9]{2}$ ]]; then
		echo "    Date format invalid (not YYYY-MM-DD)"
		return 1
	fi
	
	echo "    Date format valid, checking if date exists..."
	
	# Extract year, month, day
	local year="${date_str:0:4}"
	local month="${date_str:5:2}"
	local day="${date_str:8:2}"
	
	echo "    Parsed: year=$year, month=$month, day=$day"
	
	# Basic range checks
	if [[ $month -lt 1 || $month -gt 12 ]]; then
		echo "    Invalid month: $month"
		return 1
	fi
	
	if [[ $day -lt 1 || $day -gt 31 ]]; then
		echo "    Invalid day: $day"
		return 1
	fi
	
	# Try multiple date command formats for compatibility
	local date_error=""
	
	# Try GNU date format first
	if date -d "$date_str" >/dev/null 2>&1; then
		echo "    Date is valid (GNU date)"
		return 0
	fi
	
	# Try BSD date format
	if date -j -f "%Y-%m-%d" "$date_str" >/dev/null 2>&1; then
		echo "    Date is valid (BSD date)"
		return 0
	fi
	
	# If both fail, capture the error for debugging
	date_error=$(date -d "$date_str" 2>&1)
	echo "    Date validation failed. Error: $date_error"
	
	# For basic validation, if month and day are in reasonable ranges, accept it
	# This is a fallback for systems with problematic date commands
	if [[ $month -ge 1 && $month -le 12 && $day -ge 1 && $day -le 31 ]]; then
		echo "    Using fallback validation - date appears valid"
		return 0
	fi
	
	return 1
}

# Function to compare dates (returns 0 if date1 < date2)
is_date_older() {
	local date1="$1"
	local date2="$2"
	
	echo "    Comparing dates: '$date1' vs '$date2'"
	
	# Try GNU date first
	local date1_epoch=$(date -d "$date1" +%s 2>/dev/null)
	local date2_epoch=$(date -d "$date2" +%s 2>/dev/null)
	
	# If GNU date failed, try BSD date
	if [[ -z "$date1_epoch" ]]; then
		date1_epoch=$(date -j -f "%Y-%m-%d" "$date1" +%s 2>/dev/null)
	fi
	
	if [[ -z "$date2_epoch" ]]; then
		date2_epoch=$(date -j -f "%Y-%m-%d" "$date2" +%s 2>/dev/null)
	fi
	
	# If date commands still fail, do string comparison as fallback
	if [[ -z "$date1_epoch" || -z "$date2_epoch" ]]; then
		echo "    Using string comparison fallback"
		if [[ "$date1" < "$date2" ]]; then
			echo "    $date1 is older than $date2 (string comparison)"
			return 0
		else
			echo "    $date1 is not older than $date2 (string comparison)"
			return 1
		fi
	fi
	
	echo "    Epochs: $date1_epoch vs $date2_epoch"
	
	if [[ $date1_epoch -lt $date2_epoch ]]; then
		echo "    $date1 is older than $date2"
		return 0  # date1 is older
	else
		echo "    $date1 is not older than $date2"
		return 1  # date1 is not older
	fi
}

# Process directories
for dir in "$TARGET_DIR"/*/; do
	# Remove trailing slash and get just the directory name
	dir_path="${dir%/}"
	dir_name=$(basename "$dir_path")
	
	# Debug output
	echo "Processing: '$dir_name'"
	
	# Skip if not a directory (shouldn't happen with */ glob, but be safe)
	if [[ ! -d "$dir_path" ]]; then
		continue
	fi
	
	# Extract potential date prefix (first 10 characters)
	if [[ ${#dir_name} -ge 10 ]]; then
		potential_date="${dir_name:0:10}"
		echo "  Checking first 10 chars: '$potential_date'"
	else
		potential_date=""
		echo "  Directory name too short for date prefix"
	fi
	
	# Check if directory name starts with a valid date
	if [[ -n "$potential_date" ]] && is_valid_date "$potential_date"; then
		echo "  Found valid dated directory: $dir_name (date: $potential_date)"
		
		# Check if the date is older than current date
		if is_date_older "$potential_date" "$CURRENT_DATE"; then
			if [[ "$DRY_RUN" == true ]]; then
				echo "    [DRY RUN] Would delete: $dir_path"
			else
				echo "    Deleting: $dir_path"
				if rm -rf "$dir_path"; then
					echo "    Successfully deleted: $dir_path"
				else
					echo "    Error: Failed to delete $dir_path"
					continue
				fi
			fi
			((deleted_count++))
		else
			echo "    Skipping (date is current or future): $dir_path"
			((skipped_count++))
		fi
	else
		echo "  Skipping non-dated directory: $dir_name"
		((skipped_count++))
	fi
done

# Summary
echo ""
echo "=== Summary ==="
if [[ "$DRY_RUN" == true ]]; then
	echo "Directories that would be deleted: $deleted_count"
else
	echo "Directories deleted: $deleted_count"
fi
echo "Directories skipped: $skipped_count"

exit 0