Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
Author |
Topic |
datakeyword
Starting Member
10 Posts |
Posted - 2012-11-07 : 00:11:06
|
We often need to group the subtables during the business analytics and statistics. We know it is easy for us to group a parent table, but not that convenient to group the subtables. What your way to group a subtable? Here let’s see how SQL and esProc group the subtables. For example:To list the employee and count the cities WHERE the employee has worked over one year. Database table: staff, resume.And their main fields: Staff: nameResume: name,city,workingDaysCheck the SQL solution:SELECT name,count( *) cityCountFROM (SELECT staff.name name,resume.city city FROM staff,resume WHERE staff.name=resume.name GROUP BY name,city HAVING sum(workingDays)>=365)GROUP BY nameProcess the subtable in the way as joining the multiple tables. The grouped result set has the same number of records as the subtable. The result sets must be grouped again in order to join the records to have the same number as the primary table.Check the esProc solution: A1 =staff.new(name,resume.group(city).count(~.sum(workingDays)>=365):cityCount)Handle the subtable sets as the fields of the primary table, hence group and filter them as a regular set.So any other way to group the subtable? Welcome to discuss it with me!Author: Jim KingBI technology consultant for Raqsoft10 + years of experience on BI/OLAP application, statistical computing and analyticsedit: moved to proper forum |
|
AnalystBrown
Starting Member
7 Posts |
Posted - 2012-11-08 : 00:16:31
|
It seems that both solutions works fine with the same problems. It's just the statement is different. Except a shorter and simple statement, I don't see other advantages of your product. Why do you compare these two tools? |
|
|
datakeyword
Starting Member
10 Posts |
Posted - 2012-11-08 : 01:18:45
|
To make it easier, which will u choose? U don't need other advantages except easier and quicker. R is great, esProc is no bad too, right?Jim KingBI technology consultant for Raqsoft10 + years of experience on BI/OLAP application, statistical computing and analyticsEmail: Contact@raqsoft.comWebsite: www.raqsoft.comBlog: datakeyword.blogspot.com/ |
|
|
garyjohn
Starting Member
7 Posts |
Posted - 2012-11-08 : 02:53:19
|
Good tips! Both solutions are quite OK! But I find esProc is more suitable for business personell without too high IT technology, like marketing and sales professional, HR managers, or financial experts, esProc is good choice for them. While for upper level scientific research or big data analytics, R can not be replaced. Am I right, Jim?business analyst, marketing director |
|
|
datakeyword
Starting Member
10 Posts |
Posted - 2012-11-16 : 03:26:53
|
garyjohn, R is great, and just as you mentioned, it is not replaceable. But, every tool has its advantages and disadvantages, for exampel,r is famous for big data, esProc is outstanding for it's excell-style interface, and strong capability for handle complex structured data.Jim KingBI technology consultant for Raqsoft10 + years of experience on BI/OLAP application, statistical computing and analyticsEmail: Contact@raqsoft.comWebsite: www.raqsoft.comBlog: datakeyword.blogspot.com/ |
|
|
|
|
|
|
|