This night I was testing editing using both my Homemade testinterface and gdbi-3-pgv.jar.
Using my own testinterface I was appending a record, then I tried to get the appended record and finally I deleted the appended record and went to phpGedView and annullated all the actions taken.
Everything seemed to work okay.
Next I tried gdbi-3-pgv.jar which usually locks up when trying to add parents to a person. However, since Daniel said that this problem only happens now and then, and only happens because gdbi-3-pgv.jar looses contact with the server or the gdbi interface, I tried this several times.
Her in the early morning I discover, that my gedcom at http://phpgedview.1go.dk has increased to the double size, and completely are missing the GEDCOM END TAG.
When checking at http://eckmann.dk/phpGedView/ where I tested gdbi-3-pgv.jar many times, I found that my gedcom had increased to 15 times it's original size.
When inspecting this huge gedcom I found, that something in the GDBI interface or the gdbi-3-pgv.jar simply append a copy of the gedcom to the gedcom, and in doing that it doubles the size of the gedcom, and at the same time it looses the END OF GEDCOM TAG, replacing it with the following funny construct:
0 @F10 @I1@ INDI
Best regard
Arne
ps. In a few hours I'm leaving for a small vacation, so you won't be hearing from me the next few days.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
This is probably a problem with the gdbi_functions.php file and not with the gdbi-3-pgv.jar since all gedcom manipulation is done by the gdbi_functions.php file.
But I haven't seen this problem happen before. I have seen the loss of the 0 TRLR before, but not the duplication of GEDCOM and with the latest release the missing 0 TRLR problem should be fixed.
I reccommend starting with a fresh GEDCOM file and trying again.
I hope you have a good vacation.
--John
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Just returned home I tried to install ver. 2.52.1.
1. When clicking ZOOM in the relationship function the labels for occupation and REFN wasn't displayed, but the contents of occupation and REFN was displayed. Apparently this is only the case in the relationship function.
2. As to the GDBI problem, this was a fresh install and a fresh gedcomfile. Now, using gdbi-3-pgv.jar (UTF-8 modified) I was searching for the name: Willy Knudsen, using the Find (F3) function.
2.2. gdbi-3-pgv.jar reported that "No matching name was found"
2.3. Next I tried searching for Willy? and there I picked the name: Willy Knudsen #147 from the returned list.
3. At this point I checked and noted the size of the GEDCOM - Gedcom size = 962,317.
3.1. At this point I also checked that the "0 TRLR" was found at the end of the gedcom.
3.2. Now I picked the ADD Parents function. gdbi-3-pgv.jar asked for a name of one of the parents: I input: Faderen Knudsen.
3.3. After a wait gdbi-3-pgv.jar returned and asked for the gender. I selected "Male".
3.4. At this point gdbi-3-pgv.jar always locks up - and so it also did this time.
3.5. After a wait I again inspected the gedcom file, where I noticed that the "0 TRLR" was removed, and replaced by the following "funny" construct:
0 @F10 @I3121@ INDI
062@ FAM
1 CHIL @I47@
Now, this is a problem for several reasons: First of all, this is not correct and valid gedcom syntax. Secondly, nothing should be added to the gedcom file at this point.
3.6. At this time I force gdbi-3-pgv.jar to terminate, and then using the browser I login to phpGedView and was presented to the following messages:
Notice: Undefined offset: 2 in mypath\phpGedView\gdbi_changes.php on line 65
Notice: Undefined variable: factarray in mypath\phpGedView\gdbi_changes.php on line 69
Flgende ndringer indtraf for denne person:
Type: append
Brugernavn: myeditor
Dato: 06:09:19 26-10-2003
Gedcom: my.ged
Type: replace
Brugernavn: myeditor
Dato: 06:09:22 26-10-2003
Gedcom: my.ged
Fortryd | Vis GEDCOM Record
Knudsen, Willy
Flgende ndringer indtraf for denne person:
Type: replace
Brugernavn: myeditor
Dato: 06:09:20 26-10-2003
Gedcom: my.ged
Fortryd | Vis GEDCOM Record
Bese ndring, forskel
Flgende ndringer indtraf for denne person:
Type: append
Brugernavn: myeditor
Dato: 06:09:24 26-10-2003
Gedcom: my.ged
Fortryd | Vis GEDCOM Record
Bese ndring, forskel
3.7 Now, using undo reveals, that most can be undone, but that one can't be undone.
3.8. Now, checking the size of the gedcom revealed, that at this point the size of the gedcom file was increased from 962,317 to 1,924,656 ...
Now, these problems can be reproduced by me, over and over again, and in fact I only can reproduce these problems.
I can't say if these problems are caused by gdbi-3-pgv.jar or gdbi_functions.php. But since you and Daniel can't reproduce these errors, I think it's worth mention that my gdbi-3-pgv.jar is modified to handle the UTF-8 / ANSI situation.
In case your interested I still have the gedcom which increased to the 15,385,388 and the one which increased to 2,010,339 size, and now the latest gedcom, which increased to the 1,924,656
size.
In case your interested I can give you access to my site, so that you can reproduce the error.
best regard
Arne
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
As mentioned before, this is probably a server problem, but I wanted to let you know there is a new version of the client, gdbi-4-pgv.jar, that has the character set fix.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
In relation to this being a server problem: When I tested ver. 2.52, this problem occurred both at the 1go.dk server (Apache Server) and at the server hosting my eckmann.dk domain (Sambar server).
Thanks for the information about the gdbi-4-pgv.jar. I was planning to download the regular gdbi-3-pgv.jar and test if it behaved different to my UTF-8 version of gdbi-3-pgv.jar, but now I'll instead download gdbi-4-pgv.jar.
best regard
Arne
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
In regard to not being able to find the person, Willy Knudsen, gdbi-4-pgv.jar behaved exactly like gdbi-3-pgv.jar - i.e:
1. When using FIND to locate Willy Knudsen, it reported: "No matching name found".
2. When using FIND to locate Willy? it returned a list where the name, Willy Knudsen #147 appeared on the list.
When testing gdbi-4-pgv.jar's ability to add parents to Willy Knudsen it was found:
3. The name, Faderen Knudsen was accepted as the name of one of the parents.
4. When gdbi-4-pgv.jar prompted me for the sex of Faderen Knudsen, I clicked at Male.
5. At this point gdbi-3-pgv.jar usually locks up, but gdbi-4-pgv.jar displayed a different behaviour and displayed an error message saying:
"FATAL database error must exit"
6. At this point the "0 TRLR" line usually are removed from the gedcom file, being replaced with what I refer to as a "funny" construct, i.e:
0 @F10 @I3121@ INDI
062@ FAM
1 CHIL @I47@
This also happens with gdbi-4-pgv.jar, but this construct probably has nothing to do with gdbi-4-pgv.jar, but probably comes from a problem in phpGedViews GDBI interface.
7. When trying to undo these changes, phpGedView is unable to undo these changes, but instead phpGedView increases the size of the gedcom to approx. the double size.
As to your suggestion that these errors are created by the server, I can (again) mention, that exactly the same behaviour and errors occur at two different servers, namely a Sambar Server in Jutland, and an Apache Server here i Copenhagen.
best regard
Arne
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Oh, I see it is confusing to talk about "the server", since that usually means the http server. In the case of GDBI, you have a client, my gdbi-4-pgv.jar, and a server, John's phpGedView. So most of the problems are in phpGedView.
The problem with FIND is probably in gdbi-4-pgv.jar, so I will follow up with you off-line to get access to your database to test it.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Now, replacing the gedcom and dropping the gdbi-4-pgv.jar for a while, a tried manually to construct and apply the records using my homemade GDBI testinterface.
1. First I used the GetXref FAM last function. The function returned F1061, and therefore I retrievd F1061 and build a new record, which was submitted as an update:
2. Then I used GetXref INDI last function twice. The function returned I3121 and I3122. I therefore retrieved the two new records (one at the time) and build two new records (one at the time) which was submitted (one at the time) as updates:
0 @I3121@ INDI
1 NAME Faderen /Knudsen/
1 SEX M
1 FAMS @F1062@
0 @I3122@ INDI
1 NAME Moderen /Knudsen/
1 SEX F
1 FAMS @F1062@
These two records was accepted, and next I checked, that nothing had happened to the gedcom file from these actions.
At this point I skipped my test-interface, and using the browser I now tried to undo these chances:
3. In phpGedView I was presented to the following messages:
The following changes occured to this individual:
Type: append
Username: myeditor
Date: 05:24:21 27-10-2003
Gedcom: my.ged
Type: replace
Username: myeditor
Date: 05:41:54 27-10-2003
Gedcom: my.ged
Undo | View GEDCOM Record
View Change Diff
The following changes occured to this individual:
Type: append
Username: myeditor
Date: 05:24:33 27-10-2003
Gedcom: my.ged
Type: replace
Username: myeditor
Date: 05:42:37 27-10-2003
Gedcom: my.ged
Undo | View GEDCOM Record
View Change Diff
Notice: Undefined index: F1062 in mypath\phpGedView\functions_index.php on line 141
(Unknown) + (Unknown)
The following changes occured to this individual:
Type: append
Username: myeditor
Date: 05:25:15 27-10-2003
Gedcom: my.ged
Type: replace
Username: myeditor
Date: 05:34:45 27-10-2003
Gedcom: my.ged
Type: replace
Username: myeditor
Date: 05:47:18 27-10-2003
Gedcom: my.ged
Undo | View GEDCOM Record
Knudsen, Willy
The following changes occured to this individual:
Type: replace
Username: myeditor
Date: 05:38:11 27-10-2003
Gedcom: my.ged
Undo | View GEDCOM Record
View Change Diff
4. All the changes was undone, and at the end the gedcom was again checked and found unharmed and unchanged.
Why is it, that phpGedView accept these changes without creating havoc when the changes are made manually from my test-interface ?
When using gdbi-4-pgv.jar to try to do the same thing, the result is absolute and utter havoc.
Later today I'll try to use gdbi-4-pgv.jar from another computer, in order to check if my Java JRE perhaps is at fault, but since this happens at two very different servers I think we can safely rule the possibility out, that the problem is created by the server.
However, I'm still puzzled why changes are made to the gedcom when gdbi-4-pgv.jar is used? and no changes are made to the gedcom, when I build the records manually?
best regard
Arne
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
(the reason this name is returned is, that the name
Arne Eckmann is mentioned in a note associated with Otto Steenbergs record.)
When searching for:
Arne
phpGedView returns links to 49 records
(I guess the reason links to all these records are returned is, that the word arne probably is mentioned somewhere in these records)
When searching for:
Arne Alexander
phpGedView returns:
Eckmann, Arne Alexander Frer
(I guess the reason that "Arne Alexander" only returns one link, and that this link point to the correct person (me) is, that the combination "Arne Alexander" only occurs in my personal record).
My point here is, that the general search routine shouldn't be used to locate persons!
Now, using Soundex Search to locate Arne Eckmann, it returns a link to Eckmann, Arne Alexander Frer - i.e. the correct person.
Using Soundex Search to locate Willy Knudsen, it returns a link to Knudsen, Willy - i.e. the correct person.
My point here is, that the Soundex Search should be used to locate persons!
best regard
Arne
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Yes, normally to find persons by name you would want to use the Soundex search.
However you can do a name search in the general search using regular expressions. For example, if I wanted to search for John Finlay I could type the following regular expression into the general search box:
1 NAME.*john.*/finlay/
In my gedcom this finds John M, John Robert, and Joseph John.
To Daniel, the \s may not work coming in on a URL. When I did the search like this I got the appropriate response:
action=search&query=NAME.*char
--John
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
But I'm afraid, the usage of regular expressions in name search is too "technical" for the average - and "non-technical" - genealogy user. So using regular expression in general name search, requires the regular expressions to be added by the program.
best regard
Arne
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I believe that Daniel's program automatically builds the regular expression when it searches so that the user doesn't have to.
As you say regular expressions are too techincal for the average user. That is why I don't advertise the capability. But for the Power user it offers great flexibility in your searches. Here are some examples:
Find all people with last name Finlay who are males:
NAME.*/finlay/.*1 SEX M
Find all the Finlays who lived in Ohio:
NAME.*/finlay/.*PLAC.*ohio
You can start to see how this could become the basis for a custom report engine. Which I hope to get around to doing eventually.
--John
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Yes, I agree. And for the power user it certainly is great to have access to use regular expressions in the search box.
However, when the program automatically builds its own regular expressions, it may eventually conflict with the power users usage of regular expressions ?
best regard
Arne
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
This night I was testing editing using both my Homemade testinterface and gdbi-3-pgv.jar.
Using my own testinterface I was appending a record, then I tried to get the appended record and finally I deleted the appended record and went to phpGedView and annullated all the actions taken.
Everything seemed to work okay.
Next I tried gdbi-3-pgv.jar which usually locks up when trying to add parents to a person. However, since Daniel said that this problem only happens now and then, and only happens because gdbi-3-pgv.jar looses contact with the server or the gdbi interface, I tried this several times.
Her in the early morning I discover, that my gedcom at http://phpgedview.1go.dk has increased to the double size, and completely are missing the GEDCOM END TAG.
I only tested gdbi-3-pgv.jar once at http://phpgedview.1go.dk
When checking at http://eckmann.dk/phpGedView/ where I tested gdbi-3-pgv.jar many times, I found that my gedcom had increased to 15 times it's original size.
When inspecting this huge gedcom I found, that something in the GDBI interface or the gdbi-3-pgv.jar simply append a copy of the gedcom to the gedcom, and in doing that it doubles the size of the gedcom, and at the same time it looses the END OF GEDCOM TAG, replacing it with the following funny construct:
0 @F10 @I1@ INDI
Best regard
Arne
ps. In a few hours I'm leaving for a small vacation, so you won't be hearing from me the next few days.
Hi Arne,
This is probably a problem with the gdbi_functions.php file and not with the gdbi-3-pgv.jar since all gedcom manipulation is done by the gdbi_functions.php file.
But I haven't seen this problem happen before. I have seen the loss of the 0 TRLR before, but not the duplication of GEDCOM and with the latest release the missing 0 TRLR problem should be fixed.
I reccommend starting with a fresh GEDCOM file and trying again.
I hope you have a good vacation.
--John
Hi John
Just returned home I tried to install ver. 2.52.1.
1. When clicking ZOOM in the relationship function the labels for occupation and REFN wasn't displayed, but the contents of occupation and REFN was displayed. Apparently this is only the case in the relationship function.
2. As to the GDBI problem, this was a fresh install and a fresh gedcomfile. Now, using gdbi-3-pgv.jar (UTF-8 modified) I was searching for the name: Willy Knudsen, using the Find (F3) function.
2.2. gdbi-3-pgv.jar reported that "No matching name was found"
2.3. Next I tried searching for Willy? and there I picked the name: Willy Knudsen #147 from the returned list.
3. At this point I checked and noted the size of the GEDCOM - Gedcom size = 962,317.
3.1. At this point I also checked that the "0 TRLR" was found at the end of the gedcom.
3.2. Now I picked the ADD Parents function. gdbi-3-pgv.jar asked for a name of one of the parents: I input: Faderen Knudsen.
3.3. After a wait gdbi-3-pgv.jar returned and asked for the gender. I selected "Male".
3.4. At this point gdbi-3-pgv.jar always locks up - and so it also did this time.
3.5. After a wait I again inspected the gedcom file, where I noticed that the "0 TRLR" was removed, and replaced by the following "funny" construct:
0 @F10 @I3121@ INDI
062@ FAM
1 CHIL @I47@
Now, this is a problem for several reasons: First of all, this is not correct and valid gedcom syntax. Secondly, nothing should be added to the gedcom file at this point.
3.6. At this time I force gdbi-3-pgv.jar to terminate, and then using the browser I login to phpGedView and was presented to the following messages:
Notice: Undefined offset: 2 in mypath\phpGedView\gdbi_changes.php on line 65
Notice: Undefined variable: factarray in mypath\phpGedView\gdbi_changes.php on line 69
Flgende ndringer indtraf for denne person:
Type: append
Brugernavn: myeditor
Dato: 06:09:19 26-10-2003
Gedcom: my.ged
Type: replace
Brugernavn: myeditor
Dato: 06:09:22 26-10-2003
Gedcom: my.ged
Fortryd | Vis GEDCOM Record
Knudsen, Willy
Flgende ndringer indtraf for denne person:
Type: replace
Brugernavn: myeditor
Dato: 06:09:20 26-10-2003
Gedcom: my.ged
Fortryd | Vis GEDCOM Record
Bese ndring, forskel
Flgende ndringer indtraf for denne person:
Type: append
Brugernavn: myeditor
Dato: 06:09:24 26-10-2003
Gedcom: my.ged
Fortryd | Vis GEDCOM Record
Bese ndring, forskel
3.7 Now, using undo reveals, that most can be undone, but that one can't be undone.
3.8. Now, checking the size of the gedcom revealed, that at this point the size of the gedcom file was increased from 962,317 to 1,924,656 ...
Now, these problems can be reproduced by me, over and over again, and in fact I only can reproduce these problems.
I can't say if these problems are caused by gdbi-3-pgv.jar or gdbi_functions.php. But since you and Daniel can't reproduce these errors, I think it's worth mention that my gdbi-3-pgv.jar is modified to handle the UTF-8 / ANSI situation.
In case your interested I still have the gedcom which increased to the 15,385,388 and the one which increased to 2,010,339 size, and now the latest gedcom, which increased to the 1,924,656
size.
In case your interested I can give you access to my site, so that you can reproduce the error.
best regard
Arne
As mentioned before, this is probably a server problem, but I wanted to let you know there is a new version of the client, gdbi-4-pgv.jar, that has the character set fix.
Hi Daniel
In relation to this being a server problem: When I tested ver. 2.52, this problem occurred both at the 1go.dk server (Apache Server) and at the server hosting my eckmann.dk domain (Sambar server).
Thanks for the information about the gdbi-4-pgv.jar. I was planning to download the regular gdbi-3-pgv.jar and test if it behaved different to my UTF-8 version of gdbi-3-pgv.jar, but now I'll instead download gdbi-4-pgv.jar.
best regard
Arne
Hi again Daniel
I've now downloaded gdbi-4-pgv.jar and tested it.
In regard to not being able to find the person, Willy Knudsen, gdbi-4-pgv.jar behaved exactly like gdbi-3-pgv.jar - i.e:
1. When using FIND to locate Willy Knudsen, it reported: "No matching name found".
2. When using FIND to locate Willy? it returned a list where the name, Willy Knudsen #147 appeared on the list.
When testing gdbi-4-pgv.jar's ability to add parents to Willy Knudsen it was found:
3. The name, Faderen Knudsen was accepted as the name of one of the parents.
4. When gdbi-4-pgv.jar prompted me for the sex of Faderen Knudsen, I clicked at Male.
5. At this point gdbi-3-pgv.jar usually locks up, but gdbi-4-pgv.jar displayed a different behaviour and displayed an error message saying:
"FATAL database error must exit"
6. At this point the "0 TRLR" line usually are removed from the gedcom file, being replaced with what I refer to as a "funny" construct, i.e:
0 @F10 @I3121@ INDI
062@ FAM
1 CHIL @I47@
This also happens with gdbi-4-pgv.jar, but this construct probably has nothing to do with gdbi-4-pgv.jar, but probably comes from a problem in phpGedViews GDBI interface.
7. When trying to undo these changes, phpGedView is unable to undo these changes, but instead phpGedView increases the size of the gedcom to approx. the double size.
As to your suggestion that these errors are created by the server, I can (again) mention, that exactly the same behaviour and errors occur at two different servers, namely a Sambar Server in Jutland, and an Apache Server here i Copenhagen.
best regard
Arne
Oh, I see it is confusing to talk about "the server", since that usually means the http server. In the case of GDBI, you have a client, my gdbi-4-pgv.jar, and a server, John's phpGedView. So most of the problems are in phpGedView.
The problem with FIND is probably in gdbi-4-pgv.jar, so I will follow up with you off-line to get access to your database to test it.
Hi again
Now, replacing the gedcom and dropping the gdbi-4-pgv.jar for a while, a tried manually to construct and apply the records using my homemade GDBI testinterface.
1. First I used the GetXref FAM last function. The function returned F1061, and therefore I retrievd F1061 and build a new record, which was submitted as an update:
0 @F1062@ FAM
1 HUSB @I3121@
1 WIFE @I3122@
1 CHIL @I147@
This record was accepted by phpGedView, and nothing happened to the gedcom file.
Next I tried to retrieve this new record:
SUCCESS
0 @F1062@ FAM
1 HUSB @I3121@
1 WIFE @I3122@
1 CHIL @I147@
2. Then I used GetXref INDI last function twice. The function returned I3121 and I3122. I therefore retrieved the two new records (one at the time) and build two new records (one at the time) which was submitted (one at the time) as updates:
0 @I3121@ INDI
1 NAME Faderen /Knudsen/
1 SEX M
1 FAMS @F1062@
0 @I3122@ INDI
1 NAME Moderen /Knudsen/
1 SEX F
1 FAMS @F1062@
These two records was accepted, and next I checked, that nothing had happened to the gedcom file from these actions.
At this point I skipped my test-interface, and using the browser I now tried to undo these chances:
3. In phpGedView I was presented to the following messages:
The following changes occured to this individual:
Type: append
Username: myeditor
Date: 05:24:21 27-10-2003
Gedcom: my.ged
Type: replace
Username: myeditor
Date: 05:41:54 27-10-2003
Gedcom: my.ged
Undo | View GEDCOM Record
View Change Diff
The following changes occured to this individual:
Type: append
Username: myeditor
Date: 05:24:33 27-10-2003
Gedcom: my.ged
Type: replace
Username: myeditor
Date: 05:42:37 27-10-2003
Gedcom: my.ged
Undo | View GEDCOM Record
View Change Diff
Notice: Undefined index: F1062 in mypath\phpGedView\functions_index.php on line 141
(Unknown) + (Unknown)
The following changes occured to this individual:
Type: append
Username: myeditor
Date: 05:25:15 27-10-2003
Gedcom: my.ged
Type: replace
Username: myeditor
Date: 05:34:45 27-10-2003
Gedcom: my.ged
Type: replace
Username: myeditor
Date: 05:47:18 27-10-2003
Gedcom: my.ged
Undo | View GEDCOM Record
Knudsen, Willy
The following changes occured to this individual:
Type: replace
Username: myeditor
Date: 05:38:11 27-10-2003
Gedcom: my.ged
Undo | View GEDCOM Record
View Change Diff
4. All the changes was undone, and at the end the gedcom was again checked and found unharmed and unchanged.
Why is it, that phpGedView accept these changes without creating havoc when the changes are made manually from my test-interface ?
When using gdbi-4-pgv.jar to try to do the same thing, the result is absolute and utter havoc.
Later today I'll try to use gdbi-4-pgv.jar from another computer, in order to check if my Java JRE perhaps is at fault, but since this happens at two very different servers I think we can safely rule the possibility out, that the problem is created by the server.
However, I'm still puzzled why changes are made to the gedcom when gdbi-4-pgv.jar is used? and no changes are made to the gedcom, when I build the records manually?
best regard
Arne
At this time I've tried adding parents to Willy Knudsen using gdbi-4-pgv.jar from another computer.
Unfortunately it made no difference ;-(
best regard
Arne
Hi Daniel
Well ;-) I see that we understood different things when talking about the server ;-)
I've sent you the needed userid and pw.
best regard
Arne
I logged into your database, and I think I know why it fails when it tries to find Willy Knudsen. The GEDCOM line is:
NAME Willy /Knudsen/
It has an extra space after the first name!
John,
How do I change the query string in the search action to specify 1 or more spaces?
Hi Daniel
Unfortunately Brothers Keeper are inconsistent in the way it writes the names ;-(
best regard
Arne
Daniel,
I would use an RE like this:
String query = "1 NAME "+firstname+"\s+/"+lastname+"/";
The \s+ signifies one or more whitespace characters.
--John
The \s+ does not seem to work. I tried the following URLs to find Charlie, and only the first one worked.
action=search&query= NAME charl
action=search&query= NAME\scharl
action=search&query= NAME\s+charl
Hi John & Daniel
I have placed two dumps in this forum, which I think will reveal the nature of the problem.
The first dump is from gdbi-4-pgv.jar. The second dump is from my manual construction of the records using my homemade GDBI-testinterface.
best regard
Arne
Hi Daniel and John
When searching in my data in phpGedView for:
Arne Eckmann
phpGedView returns:
Steenberg, Otto
(the reason this name is returned is, that the name
Arne Eckmann is mentioned in a note associated with Otto Steenbergs record.)
When searching for:
Arne
phpGedView returns links to 49 records
(I guess the reason links to all these records are returned is, that the word arne probably is mentioned somewhere in these records)
When searching for:
Arne Alexander
phpGedView returns:
Eckmann, Arne Alexander Frer
(I guess the reason that "Arne Alexander" only returns one link, and that this link point to the correct person (me) is, that the combination "Arne Alexander" only occurs in my personal record).
My point here is, that the general search routine shouldn't be used to locate persons!
Now, using Soundex Search to locate Arne Eckmann, it returns a link to Eckmann, Arne Alexander Frer - i.e. the correct person.
Using Soundex Search to locate Willy Knudsen, it returns a link to Knudsen, Willy - i.e. the correct person.
My point here is, that the Soundex Search should be used to locate persons!
best regard
Arne
Yes, normally to find persons by name you would want to use the Soundex search.
However you can do a name search in the general search using regular expressions. For example, if I wanted to search for John Finlay I could type the following regular expression into the general search box:
1 NAME.*john.*/finlay/
In my gedcom this finds John M, John Robert, and Joseph John.
To Daniel, the \s may not work coming in on a URL. When I did the search like this I got the appropriate response:
action=search&query=NAME.*char
--John
But I'm afraid, the usage of regular expressions in name search is too "technical" for the average - and "non-technical" - genealogy user. So using regular expression in general name search, requires the regular expressions to be added by the program.
best regard
Arne
I believe that Daniel's program automatically builds the regular expression when it searches so that the user doesn't have to.
As you say regular expressions are too techincal for the average user. That is why I don't advertise the capability. But for the Power user it offers great flexibility in your searches. Here are some examples:
Find all people with last name Finlay who are males:
NAME.*/finlay/.*1 SEX M
Find all the Finlays who lived in Ohio:
NAME.*/finlay/.*PLAC.*ohio
You can start to see how this could become the basis for a custom report engine. Which I hope to get around to doing eventually.
--John
Yes, I agree. And for the power user it certainly is great to have access to use regular expressions in the search box.
However, when the program automatically builds its own regular expressions, it may eventually conflict with the power users usage of regular expressions ?
best regard
Arne