Hello Scott, I am a MySQL user, so you have to take what I say with a handfull of salt. You should be able to do what you descibed. If Oracle is similar to MySQL in essence, then these comments may be useful.
In the simplest INSERTions MySQL wants its data entered as complete rows. But you can direct it to place data in a particular column by using the SET column function. In the insert statement you first indicate the table and row of interest then direct its attention to a particular column. I won't trouble you with syntactical info. Another method would be to SELECT the entire row of interest and use that data to formulate an entirely new row with the new entry, then UPDATE the whole row. I have to believe that Oracle can perform one or bith of these methods. To administer the DB as a big array, you need to use an auto-incremented primary key column. For example create a column called dataID and set it to be the primary key for the table. Then every time a new row is added it automatiaclly gets a row index, starting from 0 and going up. Looking at what you have described for a table, it seems similar to the data warehousing that I do. When I perform a test I create two tables. One is for "metadata" that contains "single point" info about the test. Such as testDate, testTIME, UUT_ID, ProjectPhase, testOPERATOR. Then there is a data table containing the thousands of data points generated duriong the test. Every row of data in both tables receives the same testID for that test. The metatdata table has one row created with a testID created automatically and incremented automatically. Every row of the data table gets the same testID (as the metadata entry) and each row gets a unique dataID that is auto-incremented. You can query the metadata to find tests with similarities that you want to examine. Like, what is the testID of the tests where temperature was maintained at 150degreeC? Then you retrieve from the data table for anaylsis using those testIDs. You would have 2 arrays in the DB the row and column coordinates of which would be metatdata:(testID,column_name), and data:(dataID,column_name). Whenever I write about DB stuff I wonder if it really makes sense in the end. Good luck, Mike Ross > -----Original Message----- > From: Scott Serlin [mailto:[EMAIL PROTECTED] > Sent: Friday, January 09, 2004 2:50 PM > To: info-labview > Subject: FW: Oracle database question > > > > Can anyone tell me if it is possible to treat an Oracle > database like a giant > array? More to the point, do I have to always write a > complete record (or > row) of data into a table or is there a way to place a piece > of data into one > single column within the row and come back later and place a > different piece > of data in the same row but different column? > > Example: > I wrote the first piece of data like this: > > Table > testtime testdate dut serialnum productline > 12:00 > > > Later on I wanted to add another data point in the same row > but different > column while still maintaining the data previously entered: > > Table > testtime testdate dut serialnum productline Can I do this? Is it > allowed in Oracle? > 12:00 100100 > > Once the row is complete, I would move onto the next row. > > Table > testtime testdate dut serialnum productline > 12:00 010603 1 100100 widget > 1:00 010603 2 100101 widget > > > > > >
