|
Hello,
I'd appreciate if someone could kindly help me with the following database design issue. I've already solved it, but my solution is not much elegant. I'm working on this SQL database for one of my friends. It will make her work more efficient since she's currently dealing with a tangled up set of excel tables.
The database will hold various data for several thousand employees. This would be a piece of cake, but we also need to track the changes of those data values in time. For example, we need to know that John Smith was in Quality Department in October, but since November, he's been in Services Department, etc.
Having a separate "historical data states" table for each employee (or, alternatively, for each data type) is not a good idea, so I designed a simple table named "TimeTableItem" that stores:
- a link to the table that holds table.column names (for instance "19" for dbo.Department.DepartmentID table) to which the current row relates ("TimeTableDataTypeID" column below)
- and an ID of the related value itself (for example, "7" for Services Department within the Department table - "TimeTableDataValueID" column below)
The actual table design:
- TimeTableItemID (PK, int, not null)
- PersonID (FK, int, not null)
- TimeTableDataTypeID (int, not null)
- TimeTableDataValueID (int, not null)
- StartTime (dateteime, null) (null is treated as "from the beginning of time" 
- EndTime (datetime, null) (null is treated as "forever"
- CreatedOn (datetime, null)
- ModifiedOn (datetime, null)
Obviously, the PersonID is a link to the Person table. By the same token, StartTime and EndTime define in which time span the data for the particular person is in effect, CreatedOn and ModifiedOn denotes these self-explanatory timestamps.
The trouble with this kind of design is that we don't know beforehand to WHAT employee data (Department? Position? Salary band?) will the next row be related, so we can't set a relation between the tables here (we dont' know what will the other table be). This means that there's no way to set a foreign key constraint on a column that holds the VALUE of the related data element. This is not much consistent and might lead to a situation when my friend deletes a department from the departments table that is used somewhere else. This will result in orphaned rows.
Is there any way to solve this efficiently?
Thank you for any input,
Michal
-- modified at 21:14 Monday 4th December, 2006
|
|
|
|
|
michal.kreslik wrote: The trouble with this kind of design is that we don't know beforehand to WHAT employee data (Department? Position? Salary band?)
Then use a table for each type. A HistoricalDepartment table, a HistoricalSalary table and so on.
|
|
|
|
|
Yes, that's one of the options I was talking about, but is that really a systematical solution to implement a distinct table for each data type? I don't think so. There must be a way to "join" the individual datatype tables' values' relations in one master table.
Thanks,
Michal
|
|
|
|
|
Hello,
SQL Server 2005, VS 2005
I have a checked list box. The user will select a combination of staff members from the checked list box. This will display these staff members and their tasks that they have to do. Also I have a additional 4 check boxes on the form, where the user can select the prority to be shown. i.e from the checked list box, check staff members 1, 5, 8, 12. and display the prority as high. There 4 prority check boxes (low, normal, urgent, and very urgent)
All are dislayed in a datagrid on the form.
Currently there a 15 staff member that will be in the checked list box, but this could grow as more stafff members are added.
My problem is Writing a select query for each of the possible combinations of staff members and the prority.
I could end up writing many many select queries for each combination.
Is there a simple way to write a select query for this type of situation?
Many thanks in advance,
Steve
|
|
|
|
|
steve_rm wrote: My problem is Writing a select query for each of the possible combinations of staff members and the prority.
I could end up writing many many select queries for each combination.
Definitely DO NOT create a select statement for each possible combination. Build your select statement dynamically. When the user clicks the refresh button (or whatever), then put together a criteria list of staff member IDs and priorities.
Select * from Staff inner join StaffTask on Staff.StaffID = StaffTask.StaffID where Staff.StaffID in(1,5,8,12) and StaffTask.Priority in ('urgent','very urgent')
Be sure to check this out:
SQL Injection Attacks and Some Tips on How to Prevent Them[^]
--EricDV Sig---------
Some problems are so complex that you have to be highly intelligent and well informed just to be undecided about them.
- Laurence J. Peters
|
|
|
|
|
I'm trying to gather statistics from multiple queries to gauge our database performance. So, I dropped a "Set statistics io on" (with matching off) at the top of my query. In Sql Management Studio, the results of the query come back in the Results tab, while the IO statistics come back in the Messages tab. How can I gather the IO statistics (the messages) from .NET with the provided SQLCommand/SQLDataAdapter/etc. classes? I tried pausing my test app just after the query ran and looking through all the involved objects for the IO messages but couldn't find anything.
|
|
|
|
|
The SQLConnection object fires an InfoMessage event that you must catch.
|
|
|
|
|
Hi
Iam parsing very large text files-around 50MB each using split function. I populate a dataset from it and then use sqlBulkCopy to insert the records in the datbase. But im getting a low virtual memory error and it is taking a lot of time? Any other approaches-help?
Amna
|
|
|
|
|
Have you tried using a DTS job to do the same thing ?
Blog Have I http:\\www.frankkerrigan.com
|
|
|
|
|
No i havent. But if it is more efficient i would surely try it. Thanx
|
|
|
|
|
iam using sql server 2005 express edition. It does not have the DTS option. Any other suggestions?
|
|
|
|
|
Look at BULK INSERT in the SQL help file. If your file is in a compatible format, then this will probably be your fastest option. Otherwise, just send it one record at a time using a SQLCommand object; a stored procedure will be slightly faster than sending a bunch of INSERT statements. The DataSet object is probably taking up a lot of your memory. I routinely load text files of up to about 1GB one record at a time. I have a bunch of validation for each record, so it is very slow, but even with the validations, it loads at about 1000 records per second.
|
|
|
|
|
Sorry to bother you with this question. I'm totally new to this and I don't know what I'm doing
I want to use an Excel spread sheet as a database. Any ideas how to do it? I think I have to use ODBC but I have no idea how to write the appropriate connection string and how to write a query?
I've been googling for some time but I cannot find any information on this topic. Where should I look?
|
|
|
|
|
Search codeproject for Excel ADO. It may be better to use Access if you can, then use Excel to pull updated data from the access db.
Blog Have I http:\\www.frankkerrigan.com
|
|
|
|
|
Thank you for your advice. I've found some information that might be useful.
I want to make an application that uses a simple database that does not need SQL server, Access or MS Office installed on the computer.
I currently use Excel to do my work but I wan't to be able to do it without it.
|
|
|
|
|
You don't need MS access or office on a machine to use an access database with a website.
Blog Have I http:\\www.frankkerrigan.com
|
|
|
|
|
Using an access database is preferable to Excel, and you don't need Access or Office installed. But if you must use Excel, here is an example connection string for you:
Provider=Microsoft.Jet.OLEDB.4.0;Data Source="Path\MyExcelFilename.xls";Extended Properties="Excel 8.0;HDR=NO;IMEX=1;ImportMixedTypes=Text"
I can't remember what IMEX is, but HDR is whether the first row of data is a header row or not.
|
|
|
|
|
Thank you very much for your advice. I've found a sample on this site that seems as an excellent starting point.
I'll think about the access database too but creating an sample access database (or anyone) does require MS Access?
|
|
|
|
|
|
hi,
i have table let in the following format
Col-A col-B col-C col-D col-E
Row-1 1 2 3 4 E1
Row-2 1 2 3 4 E2
Row-3 1 2 3 4 E3
Row-4 1 2 3 4 E4
so i need the output in the format
Col-A col-B col-C col-D col-E
Row-1 1 2 3 4 E1/E2/E3/E4
i cannt able to find the way to write the SQL for it
|
|
|
|
|
I think there is no option to do this using queries.You have to go for the cursors.
Kamal
|
|
|
|
|
SELECT Col-E FROM tblSomething WHERE Col-A >= 1 AND Col-A <=4
Pablo
Sometimes I think there's no reason to get out of bed . . . then I feel wet, and I realize there is.
|
|
|
|
|
Hi, I am using ODBC to connect my VC++.NET application to connect to SQL server. I can connect any server in the LAN using server name. Is it possible to connect remote server over internet using IP address? Currently I am using connection string like
"connStr = "Driver={SQL Server};Server=server_name;Database=db_name;Trusted_Connection=yes";"
Any suggestions?
Thanks in advance.
|
|
|
|
|
Yes you can but the TrustedConnetion will likley not work. You have to use a SQL username and password.
Please bear in mind that exposing SQL server or another db to the internet is highly dangerous.
Blog Have I http:\\www.frankkerrigan.com
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
