BUG: Using Col Name Input to index_col Causes Alignment Fault (119109)



The information in this article applies to:

  • Microsoft SQL Server 4.2x

This article was previously published under Q119109
BUG# NT: 870 (4.2 - NTMIPS)

SYMPTOMS

The function index_col expects three parameters:
  • object name, where object_name is the table name.
  • index id, where index id is sysindexes.indid for the index in question.
  • key_id, where key_id is syscolumns.colid for the column in question.
This tells you the column name of a specific indexed column.

For example:
   select index_col('roysched', 2, 1) yields "title_id" as the result.
				

However, if instead of using colid as parameter three, you use a column name and if the column name is from a table containing int data, an alignment fault occurs on MIPS.

For example,
   select index_col('roysched', 2, tab1.a) from tab1
				

causes an alignment fault on MIPS.

Using a column name instead of a column id is permissible syntax and is used as an example in the "Transact-SQL Reference," page 163.

WORKAROUND

Use an integer or integer variable as the third parameter to index_col. See the SQL Server installation scripts in \SQL\INSTALL for more examples on using the index_col function.

STATUS

Microsoft has confirmed this to be a problem in SQL Server version 4.21. We are researching this problem and will post new information here in the Microsoft Knowledge Base as it becomes available.

Modification Type:MajorLast Reviewed:3/22/1999
Keywords:kbprogramming KB119109