Simplified running a command on multiple datasets

src/util/dataset-query.sh

@@ -0,0 +1,29 @@
+#!/usr/bin/env bash
+set -e
+
+# Execute parallel commands per folder, collecting the output on stout.
+#
+# USAGE
+# 	"$0" <folder(s)> -- <command(s)>
+#	"$0" "folder0" "folder1" "folderN" -- mycommand --option --whatever '{}'
+#	"$0" "folder0" "folder1" "folderN" -- echo '"\"This data from the pipe\""' '|' jq --arg folder '{}' --arg pwd '"$PWD"' "'{ folder: \$path, pwd: \$pwd, pipe: . }'"
+#
+# Separate folder(s) and command(s) with "--".
+# Use '{}' in the command to get the folder as input.
+# The entire command doesn't have to be a string - everything after '--' will be executed.
+# Quote strings used in the command twice; '"string with spaces"'.
+# Separate shell commands like semicolon, ampersands and pipe need to be quoted.
+
+folders=()
+
+while [[ "$1" != "--" ]]; do
+	folders+=("$1")
+
+	shift
+done
+
+shift
+
+command="$@"
+
+parallel --group -N 1 cd "--" "{}" "&&" "$command" ::: "${folders[@]}"