解决这个问题的一个常用方法是使用一个叫做here document的shell特性，试试这个:

One common way to solve this problem is to use a shell feature called a here document, try this:

 sqlite3 /Users/user/Documents/Test/dbName.dba <<EOS
     insert into myTable (Date, Details, Category, Average) 
               values(datetime('now','localtime'), '$1', '$2', '$3');

     select "Category1 total = " sum(Average) from (
          select * from myTable where Category = 'category1'
     );

     select "Category2 total = " sum(Average) from (
         select * from myTable where Category = 'category2'
     );

 EOS

请注意，EOS 可以是您喜欢的任何字符串(我想到的是 EndOfScript)，但它必须单独出现在文本的最后一行，并且没有尾随空格.

Note that EOS can be any string you like (I think of EndOfScript), but it must be alone on the last line of text with no trailing whitespace.

由于我不使用 sqlite3，您可能需要一些语句来关闭我不知道的批处理.另外，我不确定 '$1' 的东西会起作用，如果 sqlite3 是宽容的，请尝试使用$1"等.此外，您可能需要在 "CategoryN total = " 字符串后添加一个逗号.

As I don't use sqlite3, you may need some statment to close off the batch that I'm not aware of. Also, I'm not certain that the '$1' stuff will work, if sqlite3 is forgiving, try "$1", etc instead. Also, you may need to an a comma after the "CategoryN total = " string.

请注意，此解决方案允许您根据需要创建几乎任意大/长的 sql DML 语句.对于经常发生的事情，并且涉及大表，如果您对我们的系统有权限，您可能希望将 DML 转换为存储过程并调用它.

Note that this solution allows you to create your sql DML statements pretty much as big/long as you want. For stuff that will happen regularly and it ranging over large tables, if you have permissions on our system, you may want your DML to a stored procedure and call that.

我希望这会有所帮助.

(如果这不起作用，请编辑您的帖子以表明您正在使用的 shell、OS/Linux 版本以及您收到的错误消息的最小版本).

(If this doesn't work, please edit your post to indicate shell you are using, OS/Linux Ver and a minimal version of error messages that you are getting).