脚本小工具 - 基于Txt的简单数据表查询

[majiajun_no_33]

在Linux下写脚本的时候, 通常需要根据某些字段对数据进行检索，也就是要实现一般数据库的select功能。在脚本里面这样的功能通常要求很简单， 数据结构也不太复杂， 不需要动用真正的数据库。基于txt格式的简单表是很好的选择，但仅仅使用bash并不太好实现查询， 一般需要通过perl，python来实现。 这里给了一个例子是使用awk来完成的，大家可以直接拿去使用。

两个文件：plaindb.awk是awk的代码, simple_select是bash的wrapper. 代码见后.

以下是定义的文本数据表的格式示例:

--- nodes.conf ------

#configure test nodes here
#format: nodename IP/hostname OStype root/Admin
#e.g.: specjdriver 192.168.6.67 windows Administrator
#| nodename hostname ostype admin |
specjdriver 192.168.6.67 windows Administrator
specjsut 192.168.6.252 linux root
specjdb 192.168.6.70 windows Admin
specjemulator 192.168.6.66 linux root

其中#开始的行是注释, 会被忽略. #|...| 行定义的是字段名字 其他行为数据.

[majiajun_no_33]

下面是简单的使用例子:

[root@rac1 config]# ./simple_select nodes.conf nodename=specjdb hostname
192.168.6.70
[root@rac1 config]# ./simple_select nodes.conf ostype=windows nodename
specjdriver
specjdb

在脚本中用来做查询非常方便:)

附代码:

plaindb.awk:

------------------------------

#variable predefined:
# condition,columnids
BEGIN{
# read condition
count = split(condition,condstrs,"=");
if(count != 2 )
exit -1;
keyid = condstrs[1];
keyval = condstrs[2];
# read schema
do{
res = getline;
}while(res != 0 && $0 !~ /#\|.*\|/);
if(res ==0 ){
# print "no schema found!!"
exit -1;
}
count = split($0,schema);
colidcount = split(columnids,columnidarray,",")
for(i=1;i<=colidcount;i++)
columnindexarray[i] = -1;
keyindex = -1;
# printf "searched: |%s|%s|\n",keyid,columnid;
for(i=2;i<count;i++){
# printf "|%s| - %d \n",schema[i], i;
if(schema[i] == keyid)
keyindex = i - 1;
for(j=1;j<=colidcount;j++)
if(schema[i] == columnidarray[j])
columnindexarray[j] = i - 1;
}
if(keyindex == -1){
# print "no column found!", keyindex,columnindex;
exit -1;
}
for(i=1;i<=colidcount;i++)
if(columnindexarray[i]==-1)
exit -1;
# print "Columns found!", keyindex,columnindex;
do{
res = getline row;
if( res == 0){
# print "reached the end";
exit -1;
}
count = split(row,columns);
# print count, ":", row;
if(columns[keyindex] == keyval){
for(i=1;i<=colidcount;i++)
printf ("%s\t",columns[columnindexarray[i]]);
printf("\n");
continue;
}
}while(1);


}
/.*/{printf("");}

-----------------------------------------

simple_select:

-----------------------------------------

#!/bin/bash
CONDITION=$2
COL_ID=$3
awk -v condition="$CONDITION" -v columnids=$COL_ID -f ./plaindb.awk $1

[majiajun_no_33]

下面是简单的使用例子:

[root@rac1 config]# ./simple_select nodes.conf nodename=specjdb hostname
192.168.6.70
[root@rac1 config]# ./simple_select nodes.conf ostype=windows nodename
specjdriver
specjdb

在脚本中用来做查询非常方便:)

附代码:

plaindb.awk:

------------------------------

#variable predefined:
# condition,columnids
BEGIN{
# read condition
count = split(condition,condstrs,"=");
if(count != 2 )
exit -1;
keyid = condstrs[1];
keyval = condstrs[2];
# read schema
do{
res = getline;
}while(res != 0 && $0 !~ /#\|.*\|/);
if(res ==0 ){
# print "no schema found!!"
exit -1;
}
count = split($0,schema);
colidcount = split(columnids,columnidarray,",")
for(i=1;i<=colidcount;i++)
columnindexarray[i] = -1;
keyindex = -1;
# printf "searched: |%s|%s|\n",keyid,columnid;
for(i=2;i<count;i++){
# printf "|%s| - %d \n",schema[i], i;
if(schema[i] == keyid)
keyindex = i - 1;
for(j=1;j<=colidcount;j++)
if(schema[i] == columnidarray[j])
columnindexarray[j] = i - 1;
}
if(keyindex == -1){
# print "no column found!", keyindex,columnindex;
exit -1;
}
for(i=1;i<=colidcount;i++)
if(columnindexarray[i]==-1)
exit -1;
# print "Columns found!", keyindex,columnindex;
do{
res = getline row;
if( res == 0){
# print "reached the end";
exit -1;
}
count = split(row,columns);
# print count, ":", row;
if(columns[keyindex] == keyval){
for(i=1;i<=colidcount;i++)
printf ("%s\t",columns[columnindexarray[i]]);
printf("\n");
continue;
}
}while(1);


}
/.*/{printf("");}

-----------------------------------------

simple_select:

-----------------------------------------

#!/bin/bash
CONDITION=$2
COL_ID=$3
awk -v condition="$CONDITION" -v columnids=$COL_ID -f ./plaindb.awk $1

[tomato_potato]

很好

[formajia9]

楼主不容易，顶一下！

[treetor]

希望能看到跟多的内容啊！！！！！！！！！！！

[converf]

最近一直在充电