写这份脚本的灵感来自这本书《设计数据密集型应用》 https://github.com/Vonng/ddia
在这本书的第三章里提及了一个使用 shell 实现的数据库
#!/bin/bash
db_set () {
echo "$1,$2" >> database
}
db_get () {
grep "^$1," database | sed -e "s/^$1,//" | tail -n 1
}
笔者在这段小脚本的基础上进行了一些拓展。新增了一些功能，例如 del show 之类的。
这份脚本依赖这几个程序，可以使用类似这样的命令 ./shdb.sh test 来判断当前环境是否有这些程序
可以使用类似这样的命令 ./shdb.sh help 来输出使用帮助
这份脚本可以在 git for windows 的 git bash 里运行
<!-- 各个软件具体的版本 bash GNU bash, version 4.4.23 -->#!/bin/bash
SHELL_FOLDER=""
if [ -z "$SHELL_FOLDER" ]; then
SHELL_FOLDER=$(dirname $(readlink -f "$0"))
fi
DB_PATH="$SHELL_FOLDER/$2.csv"
version() {
echo "0.0.1"
}
help() {
echo "version: "$(version)
echo "folder: "$SHELL_FOLDER
echo ""
cat << EOF
Usage:
command [arguments]
Command:
version
help
test test dependence
create create table
example: ./dbsh.sh create default
remove remove table
example: ./dbsh.sh remove default
list show all table
example: ./dbsh.sh list
rename rename table
example: ./dbsh.sh rename old_name new_name
reindex rebuild index
example: ./dbsh.sh reindex default
set set a value by key
example: ./dbsh.sh set default key1 value
get get a value by key
example: ./dbsh.sh get default key1
del delete a value by key
example: ./dbsh.sh del default key1
truncate truncate table
example: ./dbsh.sh truncate default
show show all data for a table
example: ./dbsh.sh show default
rawshow show all data for a table with out format
example: ./dbsh.sh rawshow default
Note:
file is saved in CSV format.
set can overwrite old data.
modify this variable SHELL_FOLDER to change the directory.
key is a keyword, you can not use it.
EOF
}
db_create() {
if [ ! -f "$1" ]; then
echo "key,value" >> $1
else
echo "$1 already exists"
return 1
fi
}
db_remove() {
if [ -f "$1" ]; then
rm $1
fi
}
db_list() {
ls $(echo $SHELL_FOLDER'/*.csv')
}
db_rename() {
old=$SHELL_FOLDER'/'$1'.csv'
new=$SHELL_FOLDER'/'$2'.csv'
if [ ! -f "$old" ]; then
echo "$old does not exist"
return 1
fi
if [ -f "$new" ]; then
echo "$new already exists"
return 1
fi
mv $old $new
}
db_reindex() {
if [ ! -f "$1" ]; then
echo "$1 does not exist"
return 1
fi
sed -i '/^[ \t]*$/d' $1
tac $1 | awk -F, '!a[$1]++{print}' | tac > $1
}
db_set() {
if [ $2 == "key" ]; then
echo "set fail, key is keyword"
return 1
fi
if [ -f "$1" ]; then
echo "$2,$3" >> $1
fi
}
db_get() {
if [ -f "$1" ]; then
grep "^$2," $1 | sed -e "s/^$2,//" | tail -n 1
fi
}
db_del() {
if [ -f "$1" ]; then
sed -i $(grep -n "^$2," $1 | cut -d ":" -f 1 | sed -e "s/$/d/g" | tr "\n" ";")'' $1
fi
}
db_truncate() {
sed -i '2,$d' $1
}
db_show() {
if [ ! -f "$1" ]; then
echo "$1 does not exist"
return 1
fi
tac $1 | awk -F, '!a[$1]++{print}' | tac | cat -n
}
db_rawshow() {
if [ ! -f "$1" ]; then
echo "$1 does not exist"
return 1
fi
cat $1
}
db_test() {
test_dependence() {
type $1 > /dev/null 2>&1
if [ ! $? -eq 0 ]; then
echo "no $1"
echo "test fail"
exit 1
fi
echo " $1 ok"
}
echo "SHELL_FOLDER is $SHELL_FOLDER"
if [ ! -d $SHELL_FOLDER ]; then
echo "SHELL_FOLDER is not existent"
exit 1
else
echo "SHELL_FOLDER is existent"
fi
if [ ! -r $SHELL_FOLDER ]; then
echo "SHELL_FOLDER can not read"
exit 1
else
echo "SHELL_FOLDER can read"
fi
if [ ! -w $SHELL_FOLDER ]; then
echo "SHELL_FOLDER can not write"
exit 1
else
echo "SHELL_FOLDER can write"
fi
echo "test dependence"
test_dependence sh
test_dependence grep
test_dependence sed
test_dependence awk
test_dependence cat
test_dependence tac
test_dependence tail
test_dependence cut
test_dependence tr
echo "passed the test"
}
case $1 in
"h"|"help")
help
;;
"v"|"version")
version
;;
"create")
db_create $DB_PATH
;;
"remove")
db_remove $DB_PATH
;;
"list")
db_list
;;
"rename")
db_rename $2 $3
;;
"reindex")
db_reindex $DB_PATH
;;
"set")
db_set $DB_PATH $3 $4
;;
"get")
db_get $DB_PATH $3
;;
"del")
db_del $DB_PATH $3
;;
"truncate")
db_truncate $DB_PATH
;;
"show")
db_show $DB_PATH
;;
"rawshow")
db_rawshow $DB_PATH
;;
"test")
db_test
;;
*)
help
;;
esac
EXIT_CODE=$?
if [ ! $EXIT_CODE -eq 0 ]; then
exit $EXIT_CODE
fi
exit 0