使用teradata系统表获取列类型

Question

我想从teradata系统表（如dbc.columns）获取实际列类型。

该表具有列columntype，但它不提供实际的数据类型。

我可以输出与

select type(columnname) from table 
output: varchar2(20) 

但也有1000台和50000列。请建议一些查询，可以给我从元数据本身列的实际列类型

Answer 1

对于视图columntype将为NULL，因为它已在运行时解析。

对于单个表格，您可以执行HELP COLUMN viewname。*;来获取实际的数据类型。

顺便说一句，有没有这样的类型像VARCHAR2在Teradata的，这是甲骨文:-)

编辑： 好像你想要一个列的完整定义。这是一个SQL UDF返回你的SHOW表中找到相同的信息（如果TD14.10覆盖我不是舒尔，我也没有检查，还）：

REPLACE FUNCTION DataTypeString 
(
    ColumnType CHAR(2), 
    ColumnLength INT, 
    DecimalTotalDigits SMALLINT, 
    DecimalFractionalDigits SMALLINT, 
    CharType SMALLINT, 
    ColumnUDTName VARCHAR(128) CHARACTER SET UNICODE 
) 
RETURNS VARCHAR(60) 
LANGUAGE SQL 
CONTAINS SQL 
DETERMINISTIC 
SQL SECURITY DEFINER 
COLLATION INVOKER 
INLINE TYPE 1 
RETURN 
    CASE ColumnType 
    WHEN 'BF' THEN 'BYTE('   || TRIM(CAST(ColumnLength AS INTEGER)) || ')' 
    WHEN 'BV' THEN 'VARBYTE('   || TRIM(CAST(ColumnLength AS INTEGER)) || ')' 
    WHEN 'CF' THEN 'CHAR('   || TRIM(CAST(ColumnLength AS INTEGER)) || ')' 
    WHEN 'CV' THEN 'VARCHAR('   || TRIM(CAST(ColumnLength AS INTEGER)) || ')' 
    WHEN 'D ' THEN 'DECIMAL('   || TRIM(DecimalTotalDigits) || ',' 
             || TRIM(DecimalFractionalDigits) || ')' 
    WHEN 'DA' THEN 'DATE' 
    WHEN 'F ' THEN 'FLOAT' 
    WHEN 'I1' THEN 'BYTEINT' 
    WHEN 'I2' THEN 'SMALLINT' 
    WHEN 'I8' THEN 'BIGINT' 
    WHEN 'I ' THEN 'INTEGER' 
    WHEN 'AT' THEN 'TIME('   || TRIM(DecimalFractionalDigits) || ')' 
    WHEN 'TS' THEN 'TIMESTAMP('  || TRIM(DecimalFractionalDigits) || ')' 
    WHEN 'TZ' THEN 'TIME('   || TRIM(DecimalFractionalDigits) || ')' || ' WITH TIME ZONE' 
    WHEN 'SZ' THEN 'TIMESTAMP('  || TRIM(DecimalFractionalDigits) || ')' || ' WITH TIME ZONE' 
    WHEN 'YR' THEN 'INTERVAL YEAR(' || TRIM(DecimalTotalDigits) || ')' 
    WHEN 'YM' THEN 'INTERVAL YEAR(' || TRIM(DecimalTotalDigits) || ')'  || ' TO MONTH' 
    WHEN 'MO' THEN 'INTERVAL MONTH(' || TRIM(DecimalTotalDigits) || ')' 
    WHEN 'DY' THEN 'INTERVAL DAY(' || TRIM(DecimalTotalDigits) || ')' 
    WHEN 'DH' THEN 'INTERVAL DAY(' || TRIM(DecimalTotalDigits) || ')'  || ' TO HOUR' 
    WHEN 'DM' THEN 'INTERVAL DAY(' || TRIM(DecimalTotalDigits) || ')'  || ' TO MINUTE' 
    WHEN 'DS' THEN 'INTERVAL DAY(' || TRIM(DecimalTotalDigits) || ')'  || ' TO SECOND(' 
             || TRIM(DecimalFractionalDigits) || ')' 
    WHEN 'HR' THEN 'INTERVAL HOUR(' || TRIM(DecimalTotalDigits) || ')' 
    WHEN 'HM' THEN 'INTERVAL HOUR(' || TRIM(DecimalTotalDigits) || ')'  || ' TO MINUTE' 
    WHEN 'HS' THEN 'INTERVAL HOUR(' || TRIM(DecimalTotalDigits) || ')'  || ' TO SECOND(' 
             || TRIM(DecimalFractionalDigits) || ')' 
    WHEN 'MI' THEN 'INTERVAL MINUTE(' || TRIM(DecimalTotalDigits) || ')' 
    WHEN 'MS' THEN 'INTERVAL MINUTE(' || TRIM(DecimalTotalDigits) || ')'  || ' TO SECOND(' 
             || TRIM(DecimalFractionalDigits) || ')' 
    WHEN 'SC' THEN 'INTERVAL SECOND(' || TRIM(DecimalTotalDigits) || ',' 
             || TRIM(DecimalFractionalDigits) || ')' 
    WHEN 'BO' THEN 'BLOB('   || TRIM(CAST(ColumnLength AS INTEGER)) || ')' 
    WHEN 'CO' THEN 'CLOB('   || TRIM(CAST(ColumnLength AS INTEGER)) || ')' 

    WHEN 'PD' THEN 'PERIOD(DATE)'  
    WHEN 'PM' THEN 'PERIOD(TIMESTAMP('|| TRIM(DecimalFractionalDigits) || ')' || ' WITH TIME ZONE' 
    WHEN 'PS' THEN 'PERIOD(TIMESTAMP('|| TRIM(DecimalFractionalDigits) || '))' 
    WHEN 'PT' THEN 'PERIOD(TIME('  || TRIM(DecimalFractionalDigits) || '))' 
    WHEN 'PZ' THEN 'PERIOD(TIME('  || TRIM(DecimalFractionalDigits) || '))' || ' WITH TIME ZONE' 
    WHEN 'UT' THEN COALESCE(ColumnUDTName, '<Unknown> ' || ColumnType) 

    WHEN '++' THEN 'TD_ANYTYPE' 
    WHEN 'N' THEN 'NUMBER('   || CASE WHEN DecimalTotalDigits = -128 THEN '*' ELSE TRIM(DecimalTotalDigits) END 
             || CASE WHEN DecimalFractionalDigits IN (0, -128) THEN '' ELSE ',' || TRIM(DecimalFractionalDigits) END 
             || ')' 
    WHEN 'A1' THEN COALESCE('SYSUDTLIB.' || ColumnUDTName, '<Unknown> ' || ColumnType) 
    WHEN 'AN' THEN COALESCE('SYSUDTLIB.' || ColumnUDTName, '<Unknown> ' || ColumnType) 

    ELSE '<Unknown> ' || ColumnType 
    END 
    || CASE 
     WHEN ColumnType IN ('CV', 'CF', 'CO') 
     THEN CASE CharType 
       WHEN 1 THEN ' CHARACTER SET LATIN' 
       WHEN 2 THEN ' CHARACTER SET UNICODE' 
       WHEN 3 THEN ' CHARACTER SET KANJISJIS' 
       WHEN 4 THEN ' CHARACTER SET GRAPHIC' 
       WHEN 5 THEN ' CHARACTER SET KANJI1' 
       ELSE '' 
      END 
     ELSE '' 
     END 
; 

Answer 2

我发现这个说法

SELECT 
DATABASENAME, TABLENAME, COLUMNNAME, TRIM(COLUMNTYPE)||'('||TRIM(COLUMNNUM)||')' 
FROM (
    SELECT DATABASENAME, TABLENAME, COLUMNNAME, 
    CASE 
    WHEN COLUMNTYPE='CF' THEN 'CHAR' 
    WHEN COLUMNTYPE='CV' THEN 'VARCHAR' 
    WHEN COLUMNTYPE='D' THEN 'DECIMAL' 
    WHEN COLUMNTYPE='TS' THEN 'TIMESTAMP'  
    WHEN COLUMNTYPE='I' THEN 'INTEGER' 
    WHEN COLUMNTYPE='I2' THEN 'SMALLINT' 
    WHEN COLUMNTYPE='DA' THEN 'DATE' 
    END AS COLUMNTYPE, 
    CASE 
    WHEN COLUMNTYPE='CF' THEN COLUMNLENGTH 
    WHEN COLUMNTYPE='CV' THEN COLUMNLENGTH 
    WHEN COLUMNTYPE='D' THEN (DECIMALTOTALDIGITS||','||DECIMALFRACTIONALDIGITS) 
    WHEN COLUMNTYPE='TS' THEN COLUMNLENGTH  
    WHEN COLUMNTYPE='I' THEN DECIMALTOTALDIGITS 
    WHEN COLUMNTYPE='I2' THEN DECIMALTOTALDIGITS 
    WHEN COLUMNTYPE='DA' THEN NULL 
    END AS COLUMNNUM 
    FROM DBC.COLUMNS 
    WHERE DATABASENAME='your database' 
) TBL 

Answer 3

去通过Retrieving column and other metadata information in Teradata我建议取出从DBC表的表和列的名称，然后建立一个“简单的脚本”获取所有的信息。我对TD没有任何线索，但是从Quick Reference开始，创建一个可以动态获取结果并将其插入到自己的表中的循环应该不会太困难。

最后一个良好的情况下提出的光标=）