3

在 clojure/java jdbc 中，我了解到使用getMetaData我可以返回有关通过 jdbc 连接的数据库的大量有趣信息。这可以根据目录、模式和表名进行过滤。 

(defn get-db-metadata 
    [db-spec ]
    (with-connection (get-db-connection-map db-spec)  
      ; get columns returs the following:
      ;  TABLE_CAT String       => table catalog (may be null)
      ;  TABLE_SCHEM String     => table schema (may be null)
      ;  TABLE_NAME String      => table name
      ;  COLUMN_NAME String     => column name
      ;  DATA_TYPE int          => SQL type from java.sql.Types
      ;  TYPE_NAME String       => Data source dependent type name, for a UDT the type name is fully qualified
      ;  COLUMN_SIZE int        => column size. For char or date types this is the maximum number of characters, for numeric or decimal types this is precision.
      ;  BUFFER_LENGTH          => not used.
      ;  DECIMAL_DIGITS int     => the number of fractional digits
      ;  NUM_PREC_RADIX int     => Radix (typically either 10 or 2)
      ;  NULLABLE int           => is NULL allowed.
      ;  columnNoNulls          => might not allow NULL values
      ;  columnNullable         => definitely allows NULL values
      ;  columnNullableUnknown  => nullability unknown
      ;  REMARKS String         => comment describing column (may be null)
      ;  COLUMN_DEF String      => default value (may be null)
      ;  SQL_DATA_TYPE int      => unused
      ;  SQL_DATETIME_SUB int   => unused
      ;  CHAR_OCTET_LENGTH int  => for char types the maximum number of bytes in the column
      ;  ORDINAL_POSITION int   => index of column in table (starting at 1)
      ;  IS_NULLABLE String     => "NO" means column definitely does not allow NULL values; "YES" means the column might allow NULL values. An empty string means nobody knows.
      ;  SCOPE_CATLOG String    => catalog of table that is the scope of a reference attribute (null if DATA_TYPE isn't REF)
      ;  SCOPE_SCHEMA String    => schema of table that is the scope of a reference attribute (null if the DATA_TYPE isn't REF)
      ;  SCOPE_TABLE String     => table name that this the scope of a reference attribure (null if the DATA_TYPE isn't REF)
      ;  SOURCE_DATA_TYPE short => source type of a distinct type or user-generated Ref type, SQL type from java.sql.Types (null if DATA_TYPE isn't DISTINCT or user-generated REF)
      (into #{}
            (map #(str (% :table_name) "." (% :column_name) "\n")
                 (resultset-seq (->
                                  (connection)
                                  (.getMetaData)
                                  ; Params in are catalog, schemapattern, tablenamepattern  
                                  ;(.getColumns "stuff" "public" nil "%")
                                  (.getColumns "db_catalog" "schema_name" "some_table" "%")
                                  )
                                )
                 )
            )
      )
    )

我对这些信息感兴趣的不是数据库中的每个表，而是特定查询返回的结果集。在这一点上，我特别需要知道检索到的特定列的最大长度，例如以固定宽度格式打印到屏幕上。

到目前为止我所考虑的（我确定不理想）：

  1. 尝试解析入站 sql 语句以找出正在查询哪些表，然后专门获取这些表的元数据。事实证明，带有函数或公用表表达式等的 select 语句的结果会很复杂。我认为它很快就会变得混乱（并且不准确）。

  2. 什么也可能是基于入站查询创建一个临时视图......然后我可以在这个视图上获取元数据。但是，如果我只有与我正在使用的数据库的只读连接，这将不起作用。通常情况下，我相信我正在尝试做的事情。

  3. 获取结果，然后为返回的每一列找到值的最大长度，然后相应地创建我的固定宽度网格。如果我正在查看大型结果集，这将不会很好......

有没有更好的方法来确定我从查询中返回的所有内容的类型？其他程序如何做到这一点？似乎我应该能够从我在这里以某种方式发出查询请求时获取元数据： 

(defn fetch-results 
  "Treat lazy result sets in whole for returning a database query"  
  [db-spec query] 
  (with-connection 
    (get-db-connection-map db-spec) 
    (with-query-results res query 
      ; (get the medata here somehow for columns returned ????)
      (doall res))
    )
  )

提前致谢。

4

1 回答 1

0

您可以在 scala 中这样做，这可能会帮助您：

var stmt: PreparedStatement = null
var rs: ResultSetMetaData = null
try {
  stmt = conn.prepareStatement(query)
  rs = stmt.getMetaData()
} finally {
  cleanup(stmt)
}
}
于 2016-02-29T03:01:14.240 回答