My suggest solution (1 Viewer)

[zezo2021]

Thank you

CLIENT gave ME A set of 10 text files with ambiguous data and no field descriptions, I uploaded them and discovered that each converted file had a repeating foreign key but no primary key. The task was to join these files together.

I proposed two solutions:

1. Linking all foreign keys together and removing duplicates.

2. Adding additional fields to the join to increase data accuracy and remove duplicates.

I would appreciate your feedback on my solutions and any other suggestions you may have.

 

[tvanstiphout]

First instinct is to add additional table(s) with PKs representing the FKs.
Also talk to an SME about what the data means, so together you can assign column names.

 

[plog]

3. Have them get their crap together.

Unless you understand the real world process this data is for, you really have little hope of putting this together correctly yourself. You need someone who can know if Column 2 of File 7 is also the same data in Column 4 of File 6, etc. etc. Otherwise its bunch of educated guesses and when you go back to the client thinking you did a great job after spending tons of hours on it, they will just point out all the ways its wrong.

The real answer is make them work with you. They probably think you are the data guy, this is data, lets just dump it off on him and hell figure it out. You need to sit down with them and make them realize they need to spend a lot of time on this with you.

That may mean sitting down for half an hour with them and briefly going over everything at a top level. You going away for an hour and merging the data best you can, returning to them to have them see how it looks, going back away for an hour to do more, then back to them, etc etc.

 

[jdraw]

Further to the comments already posted, I'd like to know what questions you asked your CLIENT and what you committed to on receipt of these files.
You have almost 400 posts in this forum, I'm sure you must have agreed to something with your CLIENT.

 

[ebs17]

The task was to connect these files together.

Text files contain text. You can copy it into one file.
This results in a nice pile of data and a great connection.
Task completed.

For more qualified claims, content, structures and claims would have to be analyzed and formulated more qualifiedly for subsequent use.

 

[zezo2021]

3. Have them get their crap together.

Unless you understand the real world process this data is for, you really have little hope of putting this together correctly yourself. You need someone who can know if Column 2 of File 7 is also the same data in Column 4 of File 6, etc. etc. Otherwise its bunch of educated guesses and when you go back to the client thinking you did a great job after spending tons of hours on it, they will just point out all the ways its wrong.

The real answer is make them work with you. They probably think you are the data guy, this is data, lets just dump it off on him and hell figure it out. You need to sit down with them and make them realize they need to spend a lot of time on this with you.

That may mean sitting down for half an hour with them and briefly going over everything at a top level. You going away for an hour and merging the data best you can, returning to them to have them see how it looks, going back away for an hour to do more, then back to them, etc etc.

The client is a man of few words and has not requested a meeting or sent any questions. He said he wants results, but he does not understand what he wants.

 

[tvanstiphout]

This requires social skills before programming skills.
When you are given an assignment, you say "let me look at that, and I will get back to you if I have any questions that prevent me from completing the assignment in a timely manner (read: in a cheap manner).
Then you perform your analysis, you come to the conclusion you need more information, and you get back to him: the high-level analysis is completed, and I have questions that need to be resolved before I can continue. When do you have time for me, or who else can answer my questions?

 

[GPGeorge]

The client is a man of few words and has not requested a meeting or sent any questions. He said he wants results, but he does not understand what he wants.

In addition to the sound advice you've already received, I'd like to raise the possibility that the reason the client "is a man of few words" who doesn't want to discuss the data is that he has no idea himself what's in it and doesn't want to expose that lack of knowledge.