andrewducker (![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
andrewducker) wrote2002-09-24 07:13 pm
andrewducker) wrote2002-09-24 07:13 pm
(no subject)
Just to give you a flavour of the combination of stupidity and inspiration that makes up my work at the moment...
I'm currently midway through the task of duplicating an Excel spreadsheet that's produced on a yearly basis. There's 12 reports on this spreadsheet, all put together by hand from data in the system I produced. This being the case, it should be easier and more reliable to calculate all the figures automatically and have the system put them into Excel (Excel (and, indeed, the rest of Office) are great for this, they're entirely externally controllable, and while the programming paradigm takes a while to get your head round, it's actually very simple once you get there.
However, to get the figures in there, I first of all have to work out how the reports were produced in the first place. This is fairly simple, because they tend to be labelled quite well, but there's been a few tricky ones. Culminating in the final report that took me two days to work out. I got the right kinds of figures out fairly quickly, with a complete set of figures coming out in a couple of hours. The problem then was that my figures were too high.
This was entirely unknown. Up to this point, I'd started off with a complete set of (for instance) all the babies born in 2001, and then knocked off different chunks of them (like the ones who originated out of region, or the ones whose outcomes were unknown, or similar) until the figures matched, a process of trial and error, but rarely taking more than a couple of hours to get right. I'd never had to deal with there being too many babies.
I was dealing with the period 1999-2001, and my figures were running about 3% high. But not across the board. Some categories were 1% high, some were 5% high, so it wasn't a simple fudge factor, something in my methodology was wrong. I pulled out complete totals, to discover that my figures were seemingly correct - my total was the total number of babies born in that period. This carried on for two days, with me even sitting down with the report originator, who couldn't figure it out. Gallingly the report was based off of some figures I'd extracted for him, about 8 months ago. Even worse, I couldn't remember what I'd done to extract them.
The answer hit me about 11pm last night, as I fell asleep. I then forgot about it until this morning as I was staring at the screen trying various combinations. I tried it, it worked. I dashed through to tell the report writer about it and asked them "The list of all babies born in 1999-2001 isn't large enough. The list of all babies 'something' in 1999-2001 is. What is it?" and he instantly gave me the answer that we'd spent so long trying to find - 'treated'. Because all babies born in 1999-2001 are treated in it (save for a couple born on New Years Eve at the very end), and some babies born in 1998 are also treated in 1999. That base list was being used to extract all the data for the report, and that was what was throwing off all my calculations. We then agreed that this was a stupid way to extract the figures and that we'd use the way I'd originally extracted them, the date of birth.
And that was two frustrating days of my time.
I'm currently midway through the task of duplicating an Excel spreadsheet that's produced on a yearly basis. There's 12 reports on this spreadsheet, all put together by hand from data in the system I produced. This being the case, it should be easier and more reliable to calculate all the figures automatically and have the system put them into Excel (Excel (and, indeed, the rest of Office) are great for this, they're entirely externally controllable, and while the programming paradigm takes a while to get your head round, it's actually very simple once you get there.
However, to get the figures in there, I first of all have to work out how the reports were produced in the first place. This is fairly simple, because they tend to be labelled quite well, but there's been a few tricky ones. Culminating in the final report that took me two days to work out. I got the right kinds of figures out fairly quickly, with a complete set of figures coming out in a couple of hours. The problem then was that my figures were too high.
This was entirely unknown. Up to this point, I'd started off with a complete set of (for instance) all the babies born in 2001, and then knocked off different chunks of them (like the ones who originated out of region, or the ones whose outcomes were unknown, or similar) until the figures matched, a process of trial and error, but rarely taking more than a couple of hours to get right. I'd never had to deal with there being too many babies.
I was dealing with the period 1999-2001, and my figures were running about 3% high. But not across the board. Some categories were 1% high, some were 5% high, so it wasn't a simple fudge factor, something in my methodology was wrong. I pulled out complete totals, to discover that my figures were seemingly correct - my total was the total number of babies born in that period. This carried on for two days, with me even sitting down with the report originator, who couldn't figure it out. Gallingly the report was based off of some figures I'd extracted for him, about 8 months ago. Even worse, I couldn't remember what I'd done to extract them.
The answer hit me about 11pm last night, as I fell asleep. I then forgot about it until this morning as I was staring at the screen trying various combinations. I tried it, it worked. I dashed through to tell the report writer about it and asked them "The list of all babies born in 1999-2001 isn't large enough. The list of all babies 'something' in 1999-2001 is. What is it?" and he instantly gave me the answer that we'd spent so long trying to find - 'treated'. Because all babies born in 1999-2001 are treated in it (save for a couple born on New Years Eve at the very end), and some babies born in 1998 are also treated in 1999. That base list was being used to extract all the data for the report, and that was what was throwing off all my calculations. We then agreed that this was a stupid way to extract the figures and that we'd use the way I'd originally extracted them, the date of birth.
And that was two frustrating days of my time.
no subject
second: what, exactly, is it that you do for a living? what is your job title?
no subject
And I'm data-administrator. And a database designer/programmer/administrator. I play with data/databases until they do what they're told. Until October 28th, when I become a programmer for Standard Life, working on mainframes.