当任何 sql 失败时如何在 unix shell 脚本中停止 sqlplus 命令

Question

我是 shell 脚本的新手。 我正在为 DBA 自动化编写代码，其中我有 master.sql 文件有 n 个 sql 的声明。这些 sql 被放置在特定的文件夹中。 我有一个脚本，它正在调用这个 master.sql 并立即间接执行这些 sqls 并保存在日志文件中

sqlplus 命令正在一次保存 master.sql 中声明的所有 sql 的输出 我想要的是它必须在第一个 sql 失败时停止执行并返回失败的特定 sql 的错误或要求回滚并询问你是否要继续？ 尝试在任何时候使用 sqlplus,, sqlerror ,$?没有任何效果

sqlplus  ${USER}/${PASS}@$SID  @master.sql > /home/deploy/vidya/properties/result/result.log

现在我要做的是......当任何 sql 从一开始就失败并在命令行上提示错误并指定哪个 sql 失败时，我想停止执行 sqls

master.sql

SET ECHO ON

@/home/deploy/vidya/properties/001_dba_demo.sql
@/home/deploy/vidya/properties/003_dba_demo1.sql
@/home/deploy/vidya/properties/002_dba_demo2.sql
@/home/deploy/vidya/properties/004_dba_demo2.sql

EXIT

我的 .sh 脚本正在调用 master.sql 并且它工作正常

#!/bin/bash

echo "Enter schema Name you want to connect: "
read USER
echo "Enter password: "
read -s PASS
#echo "Enter Oracle SID for the environment you want to connect with  : "
#read SID
echo "Enter Environment for the environment you want to connect with  : "
read ENV
echo "Enter Oracle SID for the environment you want to connect with  : "
read SID

 export ORACLE_SID=${SID}

if [ $ENV ==  UAT ]
then
   echo "Connected withh UAT ORACLE HOME:"
export ORACLE_HOME=/u01/app/oracle/product/11.2.0.4/db_1
elif [ $ENV == PROD ]
then
echo "Connected with PROD ORACLE HOME:"
export ORACLE_HOME=/u01/app/oracle/product/11.2.0.4/db_1

else
 echo "Wrong Environment Entered :"

fi
export PATH=$ORACLE_HOME/bin:$PATH

echo "want to execute sql? y/n"
read yn

sqlplus  ${USER}/${PASS}@$SID  @master.sql > /home/deploy/vidya/properties/result/result.log

**

谁能帮忙？**

Answer 1

一个想法是按顺序启动 sql，然后检查它们是否有错误： 例如

sqlplus  ${USER}/${PASS}@$SID  @dba_demo.sql > /home/deploy/vidya/properties/result/result_dba_demoSQL.log
checker = $(grep "ERROR" result_dba_demoSQL.log) 
if [ test -z $checker ]
   then echo "dba_demo_SQL successfull run"
   else 
       echo "dba_demo_SQL failed"
       exit 0
fi 

exit 0部分，当grep在日志文件中发现错误时退出整个脚本。

Answer 2

我猜 sqlplus 的主要功能是立即执行 sql 并将其保存在日志文件中。它已经在做。我想要的是一个一个地执行每个 sql，如果任何 sql 失败，那么它应该停止执行并在命令行上给出错误

Answer 3

很抱歉没有使用 cmets。声望不够。 这是我在第一个答案中的解决方案。 不要通过 sqlplus 启动 master.sql，而是一个一个启动较小的 sql，并在每次执行后检查日志文件。 如果 1 号 sql 的日志文件有错误，则 if 子句中的 exit 1 将终止整个脚本。

    #First query 
    #Sqlplus command to run the first query 
    sqlplus  ${USER}/${PASS}@$SID  @dba_demo.sql > /home/deploy/vidya/properties/result/result_dba_demoSQL.log 
    #variable checker is assigned ERROR, if ERROR is found in file (you have to check what errors you can get from oracle)
    checker = $(grep "ERROR" result_dba_demoSQL.log)
    #If clause that checks if variable checker is empty or not, if variable checker is empty, it means that no string "ERROR" was found in log files, thus script continues to run; If variable is not empty (string ERROR was chatched) script terminates. 
    if [ test -z $checker ]
       then echo "dba_demo_SQL successfull run"
       else 
           echo "dba_demo_SQL failed"
           exit 1
    fi 
    # Second Query
    sqlplus  ${USER}/${PASS}@$SID  @dba_demo1.sql > /home/deploy/vidya/properties/result/result_dba_demo1SQL.log
    checker = $(grep "ERROR" result_dba_demo1SQL.log) 
    if [ test -z $checker ]
       then echo "dba_demo1_SQL successfull run"
       else 
           echo "dba_demo1_SQL failed"
           exit 1
    fi 

Answer 4

如果您的 master.sql 与您发布的示例相同，即 sql 文件本身的路径列表，您可以这样做：

    # Read the master.sql file, for every line in master.sql file 
    for i in `cat master.sql`; do
    # basename "$i" takes the name of the file from path present in master.sql 
    sql_name = $(basename "$i")
    # use path in the sqlplus command ($i from the for loop) and sqlname for the log file name
    sqlplus  ${USER}/${PASS}@$SID  @$i > /home/deploy/vidya/properties/result/result_$sql_name.log
checker = $(grep "ERROR" result_$sql_name.log) 
if [ test -z $checker ]
   then echo "$sql_name successfull run"
   else 
       echo "$sql_name failed"
       exit 1
fi ; done