扫描到无效类型时，Sqlx不引发错误

Question

I am using sqlx in Go, which is very helpful, but it does not seem to throw errors when I use struct scan and the types of the struct don't match the sql types. For example here I set up a database to have a name (string) and age(int):

+-------+--------------+------+-----+---------+-------+
| Field | Type         | Null | Key | Default | Extra |
+-------+--------------+------+-----+---------+-------+
| name  | varchar(255) | NO   |     | NULL    |       |
| age   | int(11)      | NO   |     | NULL    |       |
+-------+--------------+------+-----+---------+-------+
+------+-----+
| name | age |
+------+-----+
| bob  |  10 |
+------+-----+

I then use sqlx to read out the values into a struct, but the struct has the wrong types.

package main

import (
  "log"
  "github.com/jmoiron/sqlx"
  _ "github.com/go-sql-driver/mysql"
)

// in database name is a string and age is an int

type Person struct{
  Name int
  Age string
}

func main() {
  sqlSession, err := sqlx.Open("mysql", "root:@(localhost:3306)/dashboard?parseTime=true")
  if err != nil {
    log.Panic(err)
  }
  err = sqlSession.Ping()
  if err != nil {
    log.Panic(err)
  }
  query := "SELECT * FROM test"
  rows, errSql := sqlSession.Queryx(query)
  if errSql != nil {
    log.Panic(errSql)
  }
  for rows.Next() {
    var p Person
    errScan := rows.StructScan(&p)
    if errScan != nil {
      log.Panic(errScan)
    }
    log.Println("Person:", p)
  }
}

So instead of giving me an error, it has zeroed out values. Person: {0 }

Has anyone else run into this problem? Does any one else think this is a bug? I think it should give me an error when I try to scan into an invalid type.

Answer 1

I suspect that is a side-effect of deciding not to return errors when it can't find a place to put the returned value. One reason you may not want to return an error is if you do a "select *", then even if your struct has all columns, you will not be able to add new columns without the code immediately beginning to return errors.

One could argue that if all the query columns were unused, then it should return an error. Probably opening an issue with the project would be the way to go.