|
Having done this sort of thing in the past (and yes it was for the banking industry) you are going to need someone with domain knowledge, making an incorrect assumption on the relevance/relationship of a column can lead you down some nasty cul de sacs.
Never underestimate the power of human stupidity -
RAH
I'm old. I know stuff - JSOP
|
|
|
|
|
Delete a column, see who complains and then get them to explain what it's for.
// TODO: Insert something here Top ten reasons why I'm lazy
1.
|
|
|
|
|
see if you can get an input screen (or a few) and some reports and open/run it for a specific record. Next you need to try and see if you can match the data for a specific record to fields on the input screens or reports. That will give you a good understanding of how some of the fields fit together.
|
|
|
|
|
Southmountain wrote: what is the best way to get understanding of this table quickly?
There isn't one. There are tools and scripts for most DBs that can create some type of analysis for you, but you really don't need to understand the fields, you need to understand the data.
We can all grouse and speculate about the "100s of fields", but let's assume there is a valid reason for them even though I'm hard pressed to come up with one.
What type of understanding are you trying to achieve? Data is data and the question is if and where it is used. I'd suspect there could be a lot of drop columns in your future, but that requires a detailed look at your recordset objects in the code that is using the DB. Honestly, it's a flat table so despite the crazy column count, it should be clear to understand. If it has a bunch of relations, that could take a lot of caffeine or alcohol, or both 
What's your scope of work in relation to this monster? Crazy as it seems, the DB could be oddly efficient depending on the use of the data. You know, Select * (perish the thought!) from tablename where id=x is pretty simple, lol. If your task is to clean up and reduce the size of the database, that is one thing. If you're stuck with it, it is what it is and how the data is used is of the utmost importance.
|
|
|
|
|
Understand ‘quickly?’: probably not possible.
Step back. Look at the application and the interfaces that update the table. Depending on the database tech, there’ll be a way to search procedures for the table name. Study these procedures.
Then (or while doing the above) look at a subset of the data, such as the last day’s worth of records.
Good luck. 
Time is the differentiation of eternity devised by man to measure the passage of human events.
- Manly P. Hall
Mark
Just another cog in the wheel
|
|
|
|
|
This mostly applies if it's a relational table. There a lot of excellent responses here, but one topic that I didn't see explicitely mentioned was "Normalization".
Normalize the data. You might need to have a sample of data to work(play) with to understand.
There are better resources to explain normalization further, and *some* SQL books hit the topic early (the good ones, anyway).
Normalization is at the root of what many here have indicated are problems - like too many columns in individual tables! You likely have key fields and data fields repeating all through your data. Even if you can't change the tables, understanding how it isn't normalized will help you understand the data better. (Maybe a future task will let you fix this)
> You'll need to understand the domain. It would be great if you have some people that can get you pointed in the best area. It sounds like there is a lot more complexity than can be dealt with quickly.
> Maybe there's a program that uses that table that might reveal columns content or purpose?
> Review a small sample of data in a dev/work database.
> For longer-term goals
>try to use the right terms too, especially when consulting your database expert. e.g. It might be habit to say 'fields', but in the relational database world it's a 'column'. Stuff like that will make communication more consistent.
> Maintain a group of documents, references for yourself of what you're finding - and share it.
> Topics and terms to be familiar with, or review:
> the different between a data table versus a reference table
> A Many-to-Many relationship and a "Join Table" (sometimes called a "Pin Table" etc. ...but not by me)
I didn't intend to write a book here! You have too much to read here already! I wish you success!
|
|
|
|
|
Others have mentioned researching input screens/forms and reports.
If you can capture the queries/views/procedures from the database side that are involved that could help out.
|
|
|
|
|
I work for a finance institution, and our tables are highly un-normalized with all the problems this incurs.
we have many tables with hundreds of columns.
But people in finance are used to excel like spreadsheet and they want to see everything at one place.
I came to that place with a background in databases and I was horrified and still am.
But it is hard to change legacy code AND mentalities. So for better or worse I'm stuck with it.
The best advice I can give you: find the queries people/programs make the most often, this often reveal the «subtables» within the big one.
Also if you have self join that often indicate two tables (or more) sharing the same table data.
Once you have that draw a diagram as if those tables really exist.
If you can create views that represent those tables.
This way you will better understand the «schema»
Quote: «Il faut imaginer Sisyphe heureux»
Albert Camus
|
|
|
|
|
the engineer job description. I loved it.
Also the thought that jumped in my head.
Data Scientist.
Forecasts the future using neither data nor science.
See also, Weather Forecaster
Project Manager
Not a Manager of people nor do they have only one project
To err is human to really elephant it up you need a computer
|
|
|
|
|
data scientist is a fancy term now...
diligent hands rule....
|
|
|
|
|
I bet you really like the titles:
1. DevRel (Developer Relations)
2. Developer Advocate
Here's the meme I created for DevRel (snapshot)[^]. 
|
|
|
|
|
Development Manager:
Safety device that operates to provide deadline-overpressure protection of a VP and CEO. If project fails because of the high circle's stupidity, those melts.
Advertise here – minimum three posts per day are guaranteed.
|
|
|
|
|
Yes, but ... weather forecasts use heavy computer models etc
it seems to me that a lot of real science is in there actually 
so this metaphor could be a bit stretched ...
|
|
|
|
|
They are calling me (phone) and sending me e-mails every day.
Opening their first mail using Notepad++ and removing all \0 chars, I've seen they are trying to make me participate in a free training course to learn how to sell their products better.
A guy calling from Washington USA told me I have been so good selling their products I deserve to participate in their PARTNER SALES ACCELERATION SERVICES!
I declined his invitation during that first call, but they keep calling and sending mails.
I HAVE 2 OFFICE LICENCES!!!
 ![Elephant | [mastadon]](https://codeproject.global.ssl.fastly.net/script/Forums/Images/elephant.gif)
|
|
|
|
|
Are you sure that it is Microsoft, and not some scammer trying to get your money? The latter being much more likely.
|
|
|
|
|
Why can't it be both? 
"These people looked deep within my soul and assigned me a number based on the order in which I joined."
- Homer
|
|
|
|
|
Mail servers seem legit and they don't ask me to pay anything, they want to teach me how to sell better their products...
The number they are calling me from looks legit too: 12064299176.
In any case, it looks like phone numbers can be spoofed too, them all.
|
|
|
|
|
Not sure about the phone number. Looks like a scam to me. The "toll free" extension in US is 800, not 206.
Advertise here – minimum three posts per day are guaranteed.
|
|
|
|
|
Joan M wrote: The number they are calling me from looks legit too: 12064299176.
According to this, it's a scam center using a Seattle number.
Jeremy Falcon
|
|
|
|
|
What you’re not saying is if you did indeed take part in their sales program. The question here is how did they come into possession of your email address and phone number.
|
|
|
|
|
Nope I did not.
They have my name, my mail address, and my phone number.
And I am not aware of having that information anywhere available.
Just proceeded in blocking their phone number and marking their mails as spam. Yesterday I got 3 calls while I was speaking with a customer.
|
|
|
|
|
It's not Microsoft. It's a third party that has likely purchased your information from the dark web. I guarantee everybody's email, name, address, phone, mother's maiden name, pet's first name, etc., is available for sale on the dark web.
|
|
|
|
|
Just noting that the Microsoft does have a Sales partner program. They have had it for probably 20 years at least. Basic idea is that the partner has some sort of service/product that depends on Microsoft and so the partner can do a sell through and the partner does (or at least did) get a discount on products/services which they can then sell to the partner's customer.
Sell with Microsoft[^]
And I am rather certain they promote it.
So quite likely that someone could be trying to make a quota by getting people to sign up for it.
|
|
|
|
|
yes, maybe... then that someone has got blocked.
Everything looks legit (I know I can't trust that), if it's not legit I am amazed they have all the details of the company + phone + my name... that's impressive. If it's legit... I am amazed because of how they act.
The most surprising part is that MS will have sold gazillions of licenses... how come they can call me 1237894623 times a day to help me sell better their products... Now that I've written it... maybe I am the guy with less MS licenses in the planet...
|
|
|
|
|
Email's easy to block, fortunately. But phone number?
Next time they call, tell them you have a new phone number.
Then give them the number for some other internal department within MS. Or your worst enemy's.
If they call back at your previous number after that, stop answering. Presumably they call from the same number?
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
