Merging problem?

[Neil Grantham 0]

I've read Terry's website about Merging, but still can't see why my merge went wrong

 

I had my own Project with just one data set.

I imported a GEDCOM which created a second data set

 

I wasn't too fussed which would remain as the data set after a merge and delete, but:

 

If I did B to A than deleted A I lost 'my' original data

If I did A to B and deleted B it worked.

 

I'm not certain it was this way round but certainly one way I lost data but the other it didn't.

 

Why would this be, surely, which ever way you merge, after deletion the result should be much the same?

 

Using 6.12 UK Gold

[Terry Reigel 0]

I wasn't too fussed which would remain as the data set after a merge and delete...

Then I'm afraid you missed the point of the article.

 

Why would this be, surely, which ever way you merge, after deletion the result should be much the same?

No, it certainly is not. That's the whole point of the article.

 

A "data set merge" appends the contents of one data set to the other. When you do that, one data set remains unchanged, and the other has the contents of what was formerly in both of them. Obviously, if you then delete the one that has the combined contents you will lose data. So which one you delete matters, a lot.

[Neil Grantham 0]

Terry

 

I hear what you are saying, and indeed the contents of your article, but, is it not true that if I do the reverse actions in two merges, IE

A to B and delete B or

B to A and delete A

should it not produce a Data set with the same data. Surely?

 

By definition I am merging data from one into the other, so in either method, I should have the same amount of records.

The result of what I did is that after doing it one way, I lost 'my' original dataset, the other way it worked.

I know I did the exact opposites. I restored an old project when it went 'wrong' and then re-imported the Gedcom so that I was doing the same things to the same data prior to the merging.

I wrote down what I was doing to make certain I was deleting opposites to the previous attempt

 

So, if it works one way, why not the other.

 

Or... Am I just missing a point somewhere?

 

To reiterate

I had a dataset called COOK, I imported a GEDCOM, so lets call the datasets COOK and GEDCOM. Dataset manager gave them an ID of 2 and 3 respectively.

 

First attempt, I merged GEDCOM in to COOK and deleted GEDCOM, resultant dataset still called COOK, but data was lost (my wife's family)

Second attempt, I merged COOK in to GEDCOM and deleted COOK, resultant dataset called GEDCOM. Data is all there.

[Virginia Blakelock 0]

Terry

 

I hear what you are saying, and indeed the contents of your article, but, is it not true that if I do the reverse actions in two merges, IE

A to B and delete B or

B to A and delete A

should it not produce a Data set with the same data. Surely?

 

By definition I am merging data from one into the other, so in either method, I should have the same amount of records.

The result of what I did is that after doing it one way, I lost 'my' original dataset, the other way it worked.

I know I did the exact opposites. I restored an old project when it went 'wrong' and then re-imported the Gedcom so that I was doing the same things to the same data prior to the merging.

I wrote down what I was doing to make certain I was deleting opposites to the previous attempt

 

So, if it works one way, why not the other.

 

Or... Am I just missing a point somewhere?

 

To reiterate

I had a dataset called COOK, I imported a GEDCOM, so lets call the datasets COOK and GEDCOM. Dataset manager gave them an ID of 2 and 3 respectively.

 

First attempt, I merged GEDCOM in to COOK and deleted GEDCOM, resultant dataset still called COOK, but data was lost (my wife's family)

Second attempt, I merged COOK in to GEDCOM and deleted COOK, resultant dataset called GEDCOM. Data is all there.

 

You had a dataset called Cook.

You imported a dataset called Gedcom

You said the dataset manager gave them IDs of 2 and 3.

Was dataset 1 involved at all?

 

From Help:

"NOTE: If you merge B to A, A will contain both data sets and B will remain the same as it was before the merge. The ID Numbers for individuals in A will stay the same and the individuals in B will be appended to the end of A. If you merge A to B, B will contain all the data and A will remain the same, etc. Wholly Genes, Inc."

 

A to B, or B to A: the first letter is the dataset that remains intact; the second letter is the merged dataset.

 

If Cook is A and Gedcom is B, say you merge B to A. You are copying the people in B-gedcom to A-Cook. The people in A-Cook keep their original person ID numbers and you add the B-gedcom people on to the end with new person ID numbers. You only copied the people in B, so you now have 2 datasets in the project: Cook+Gedcom AB and Gedcom B. If you are very sure the merge went the way you wanted, only then would you delete the standalone dataset (B-gedcom in this case).

 

No data is lost after the merge. No matter whether you merge A to B, or B to A, you end up after the merge with an AB dataset and an original B or A dataset. The dataset to which you append the other dataset retains the original person ID numbers.

 

After the merge, if you are not sure which dataset is which, open the simple Picklist, filter by each dataset in turn to see how many people are in each dataset and which have retained their original person ID numbers. As long as you don't delete the merged dataset in error, I don't see any way to lose data.

 

I think it's important to consider the consequences of the two merge options. Maybe you'll decide to keep your original dataset intact, i.e. merge A to B (in case you don't like the merge results), but then you forfeit your original A person ID numbers in the merged dataset. Always back up the project before you merge!

 

Virginia

[Neil Grantham 0]

Virginia

 

I may have got the notation of the A to B wrong, but my explantion using COOK and GEDCOM is how I did it.

I most definitely followed what I thought to be the procedure and definitely in one case, after the merge and delete, I lost the data from my wife's dataset (COOK) whilst the other way round it worked and I am now left with the GEDCOM dataset (3)

 

The reason I didn't mention dataset 1, this is the result of a similar exercise some months ago, where I seem to recall the same happened and I ended up with dataset 2!

Now I'm on 3 !!

 

I'm going to repeat this again with my saved backup of the original project (pre-gedcom), both ways to see if I'm just daft

Edited January 10, 2007 by Neil Grantham

[Virginia Blakelock 0]

I'm going to repeat this again with my saved backup of the original project (pre-gedcom), both ways to see if I'm just daft

 

 

You're not daft - just have to be very sure which dataset you are deleting. If you are happy with the merged dataset and have no use for the intact dataset, you can delete the latter. If in doubt, make the intact dataset inactive in the Data Set Manager until you are positive. In the screenshot below you can see how each option determines which is the merged dataset (contains both datasets).

 

Virginia