Peter's Address Summary Report

YouTube · Post by **Jane** » 28 Feb 2012 08:26

Peter a couple of things.

1. If you click the X on the prompt the plugin runs
2. I don't think you need the 'Cancelled' warning when you hit the cancel warning.
3. What would help me as I only store the bit if the address which is not in the Place would be an option to append the Address data to the Place when running the report.

ID:5964

PeterR · Post by **PeterR** » 28 Feb 2012 16:49

Jane,
Many thanks. I appreciate (and hope I learn from) your feedback.
(1) Oops! Version 1.1 seems to handle Close (and Escape) better.
(2) I agree - removed from V1.1.
(3) I do exactly the same, and have added, in V1.1, an option for appending Place (if present) to each Address.

YouTube · Post by **Jane** » 28 Feb 2012 17:12

Thanks Peter,

That's approved now.

Another couple of Ideas.

What about an option to list all the Facts which relate to the address, so not combining in the addresses (if you use an item pointer in the column the live clicking from the result set will work).

and. . .

A filter search so you could find for example all the addresses with Liverpool in them somewhere.

PeterR · Post by **PeterR** » 28 Feb 2012 22:30

Thanks, Jane.
I've thought about your further suggested options, and I wondered, first, might this be better as a separate plugin, since it would no longer be a 'summary' and would need different columns, etc. Then I wondered, second, whether this would have any real advantages over a Fact Query - I already have one that does more-or-less what you suggested and could be modified to restrict the scope of =ContainsText() to the Address (and Place, of course). Please see earlier thread:
http://www.fhug.org.uk/cgi-bin/index.cg ... num=4550#7
Horses for courses? Please let me know.

Post by **tatewise** » 29 Feb 2012 13:11

Peter, your Plugin works very well, and the following are comments on possible but non-urgent areas to review.

I have some quite varied Address entries as summarised below.

(1) Complex Numbers
These have their leading number separated, but leave the suffix letter or other symbols with the rest.
Perhaps a letter suffix could be included with the number if there is no space character in between.
Perhaps all non-alpha suffix characters could be included with the number.

21a Fann Street, Aldersgate, London
12b Hanworth Road, Hounslow
35/1 Allen Street, Sheffield
4 / 6 Chester Street, Sheffield
8 / 39&40 / 54 / 534 Broad Street, Banbury
39 & 40 Broad Street, Banbury
8 Broad Street, Banbury

Interestingly, if Ignore leading numbers is chosen, all three Broad Street, Banbury addresses are initially sorted together, until I click the Address column heading, which sorts them according to the symbols!
So it seems that the fhOutputResultSetColumn sort criteria are different from the Result Set criteria.

(2) Numbers with Alpha Prefix
These have no leading number, but would sort better on the street/road if all the alpha prefix before the number were considered part of the number.
Perhaps if any comma separated part of an Address starts with a number (ignoring spaces), then all preceding parts could be treated as a prefix on the number.

Flat 1, 144 Moscow Drive, Liverpool
Golden Ball, 26 Campo Lane, Sheffield
Goodrington Methodist Church, 1 Grange Road, Paignton
Greenbanks, 5a Greenway Road, Galmpton, Brixham
No.1 Court 11, Dorset Street, Sheffield
No.4 Court 1, Bethel Walk, Sheffield

(3) Newline Blob Symbols
For consistency with FH, it might be useful to represent each newline character with a blob symbol.
The blob is the ASCII Bell code 7 e.g. string.char(7)

(4) Reverse Address Sort
This has been mentioned on the FH Mailing List.
The idea is to sort each column part, but giving the right-hand columns priority.
Thus all addresses with the same Town in right-hand column would come together.
If the Address+Place option is used then all with the same Country in right-hand column would come together.
This approach also side steps the problem with leading numbers, which would get sorted with lowest priority.

I am happy to advise on any tricky LUA scripting if necessary.

PeterR · Post by **PeterR** » 29 Feb 2012 15:03

Thanks, Mike. I also have quite a few 'tricky' addresses, but decided to quit while I was ahead. There are some issues that would be quite an effort to resolve in your areas 1 & 2:

(1) Complex Numbers
I was keen to achieve numerical sorting of numbers as far as possible, and including letters and/or additional numbers in the Number column raises philosophical as well as practical problems.

The displayed Address column is not used for sorting. There is an almost hidden (width = 1) column containing the Address with leading non-alphabetics removed, which is sorted.

(2) Numbers with Alpha Prefix
Sorting such addresses numerically would be as difficult as in (1), above, but I agree that the Ignore Leading Numbers option would work better if the whole number-containing prefix were ignored for sorting. The big problem I foresee here is addresses which contain a postcode (Zip code, etc.). Such codes have a variety of formats and could occur in addresses without a building number, and need not be the last part of an address.

(3) Newline Blob Symbols
FH5 now displays multi-line addresses as if single-line with comma and space separating the parts, in all reports (except, quite correctly, the Record Detail reports) and I thought I would do the same in this plugin. I suppose the blob symbol could indeed be used instead, with no loss of readability, but the Place parts (if present) would be separated by commas anyway. And what symbol should be used between the last Address part and the first Place part?

(4) Reverse Address Sort
I noticed this in the FH Mailing List, and provided that option in the Places Summary Report plugin. It should be possible to do likewise as an extra option for the Ignore Leading Numbers option.

PeterR · Post by **PeterR** » 01 Mar 2012 17:43

V1.2 has implemented Mike's suggested items (3) and (4). I decided on ' / ' as separator between Address and Place. I managed to maintain as much numeric sorting as possible when Address (and Place) parts are reversed. For area (1), I used an additional 'hidden' column for the numeric sorting, which improves the display of some of the 'problem' numbers.

I've not yet figured out if it will be possible to deal with all possible alphabetic prefixes for numbers without causing more problems than it solves.

Post by **tatewise** » 01 Mar 2012 22:15

Both the Address Summary and Place Summary are really very good.

Once FH V5 is available, so I can organise all my Address data into consistent designated Column Parts, the same as I have for Place Column Parts, I believe they will work well.

The few 'unusual numbers' that don't quite work as I would like, can be easily adjusted to fit the Plugin.

You don't need to worry about alphabetic prefixes, because as I suspected, with the reverse sort, they drop into place correctly.

I have one small comment on the Address Summary two forward sorts.
It looks as if you could make the 'hidden' Address column the genuine Address column, and make the No. column hold everything that prefixes its Address entry.
Then clicking on the Address column heading will sort OK.

The same is almost true for the Address Summary reverse sort.
The differences between the Address / Place - Reversed and the hidden column sort order is not significant.

BTW: The character I chose to separate Address and Place in my Map Life Facts is the triple dot ellipsis ().

PeterR · Post by **PeterR** » 01 Mar 2012 23:12

Thanks for your kind remarks, and support. I have quite a few house names in quotation marks, and that persuaded me to the present method.

The present reversed sort puts the addresses into street order within locality, and then proper numerical order within street. Clicking the Address / Place - Reversed column heading puts the addresses in alphanumeric order of number + street, which I find less appealing. I suppose it depends how many addresses you have in the same street and in different streets in the same locality.

I've enjoyed learning a few features of Lua, anyway.

BillH · Post by **BillH** » 06 Mar 2012 01:49

Peter,

Really like the report your plugin generates, but I'm getting some odd results. Here is an example.

Note that the actual address for the second entry showing on the report is '1016 3rd St' and for the fourth one it is '1105 90th St SW #2 98204'.

The numeric portion of the street name for these is ending up in the 'No.' column rather than the 'Address / Place' column. Shouldn't the street name be in the 'Address / Place' column as it is for most other addresses?

Then, if there is no number, just a numeric street as in '12th Street', it ends up like this:

Here again the street number is getting put into the 'No.' column rather than the 'Address / Place' column, but this time even the alpha portion of the street is being put there.

Is this the way it is supposed to work?

Thanks,

Bill

Post by **tatewise** » 06 Mar 2012 12:11

This illustrates the complex nature of Addresses.

Peter, I have been experimenting with an alternative technique.
It involves splitting each Address/Place into separated Column Parts.
These are then presented in the Result Set with each Column Part in its own Result Set column.

For normal sorting, the columns are left justified, with empty columns on the right.
Each Column, left to right, has an ascending iSortPriority from 1 to max number of columns.

For reverse sorting, the columns are right justified, with empty columns on the left.
Each Column, left to right, has a descending iSortPriority from max number of columns to 1.

It basically seems to work OK, but I have not incorporated the processing of leading numbers yet.

If users have organised the Address and Place data into consistent columns, then its works even better.

PeterR · Post by **PeterR** » 06 Mar 2012 16:38

I used to work for Royal Mail so am all too well aware that a valid UK postal address can contain anything from three to eight lines (not including country).

I can certainly see the appeal of a separate column for each Address and Place part, but standardized use of particular columns is unlikely. Currently, I find the distinction in the results between Address parts and Place parts quite useful in helping me to be consistent at lest as far as which parts belong in Place and which in Address.

Leading numbers were always going to be tricky, especially with UK house numbers often featuring an alphabetic suffix or more than one group of numeric digits, while US addresses often feature a thoroughfare name which begins with a numeric digit. And continental Europe address formats are different again.

BillH · Post by **BillH** » 06 Mar 2012 16:49

Peter,

Seems like it is a tough thing to handle all situations.

Would it be possible to have an option to put the entire address back together in the report rather than separating out the possibly incorrect 'No.' part into a separate column? I realize you probably need this in order to sort the data into the correct order, but it really doesn't need to be split out in the report.

It just looks funny to have the incorrect 'No.' portion broken out that way.

Bill

PeterR · Post by **PeterR** » 06 Mar 2012 17:20

Bill,
Thanks. I'm trying a few possible improvements, and your suggestion of an option to output only the complete string of parts may be the best.

PeterR · Post by **PeterR** » 06 Mar 2012 22:20

V1.3 seems to deal with Bill's ordinal numbers OK, and everything else still seems OK for all the addresses I've tested. I also fixed a bug where I hadn't been rigorous enough in the coding when moving on from processing a Place.

BillH · Post by **BillH** » 07 Mar 2012 17:57

Peter,

I tested version 1.3 and it does work with all my addresses now.

Thanks!

Bill

Family Historian User Group