* Sort order for Place Records

[AdrianBruce]

Background:
I have been promoting some of my locations from Addresses to full Place Records. This is so I can map churches, hospitals, cemeteries, etc. (My mapping - such as it is - is driven from the Place Records). Most of my churches have a dedication, e.g. "St. Matthew's, Haslington, Cheshire, England" so the Place Records sort in an obvious manner, and the type-ahead / predictive text works in an equally obvious manner (i.e. "St. Mat" may suggest the correct place.)

Non-conformist chapels tend to pose more of a problem. Some have similar dedications while others have names that work equally well for sorting and type-ahead (e.g. "Hope Congregational Chapel, Clifton, Gloucestershire, England").

The rest can be a problem when there is no specific name apparently associated with the chapel, only a description, e.g. the Wesleyan Methodist Chapel in Tunstall, Staffs. If I enter the Place Name as "Wesleyan Methodist Chapel, Tunstall, Staffordshire, England", then my narrative sentences read like "She was baptised in Wesleyan Methodist Chapel, Tunstall ...". ("Baptised at ..." reads equally poorly to me). I can alter individual narrative sentences to read "She was baptised in the Wesleyan Methodist Chapel, Tunstall ..." (italics added here for highlighting) but I dislike altering too many narrative sentences as any subsequent change of the template for the sentence is not reflected in the altered sentences.

My current solution is to alter the Place Name to "the Wesleyan Methodist Chapel, Tunstall, Staffordshire, England" (italics added here for highlighting again). However, this can be clunky when entering new events if I don't remember whether I entered the Place as "the Haslington Chapel..." or "Haslington Chapel...", say.

Proposal:
Place Records should be provided with an optional "sort order name". (I would enter "the Haslington Chapel..." as the Place Name, so it is used in narrative sentences and enter "Haslington Chapel..." as the sort name).

Then the Places tab should show the places in an order defined by the sort order name (if present) or the normal name. Also, when entering or updating places against facts (events, attributes, etc.) the type ahead facility should look at both the normal name and the sort order name, i.e. if I type "Haslington Cha..." it suggests "the Haslington Chapel..." and if I type "the Haslington Cha..." it also suggests "the Haslington Chapel...".

[tatewise]

As a workaround consider this:
Always make "Haslington Chapel, ..." the Place name, so sorting and type ahead works as normal.
In the Place record local Note enter "Article: the " to specify when the definitive article is needed.
In the Sentence Template, use {=GetLabelledText(%FACT.PLAC>NOTE2%,"Article:")} where the definitive article may appear.
This is a variant of the FHUG KB Narrative Report Fact Sentence Templates Custom Fact Fields suggestion.

[davidf]

The rest can be a problem when there is no specific name apparently associated with the chapel, only a description, e.g. the Wesleyan Methodist Chapel in Tunstall, Staffs. If I enter the Place Name as "Wesleyan Methodist Chapel, Tunstall, Staffordshire, England", then my narrative sentences read like "She was baptised in Wesleyan Methodist Chapel, Tunstall ...". ("Baptised at ..." reads equally poorly to me).

What about "Tunstall Wesleyan Methodist Chapel, Tunstall, Staffordshire, England"?

I put "York Minster, York, Yorkshire, England", or "Bamburgh Castle, Bamburgh, Northumberland, England", so is there anything wrong with repeating the place name, if you don't want "The Minster, York ..." or "The Castle, Bamburgh ..."

What would you do with a baptism in the Cathedral at Durham?

"Durham Cathedral, Durham, County Durham, England"
or
"The Cathedral, Durham, County Durham, England"
or
"The Cathedral Church of Christ, Blessed Mary the Virgin and St Cuthbert of Durham, Durham, County Durham, England"?

I'd probably go with the former.

I would like some form of syntax to abbreviate when displaying and printing (but not when sorting) the likes of:
Oxford, Oxfordshire
Gloucester, Gloucestershire
Worcester, Worcestershire
Durham, County Durham
etc.
However I suspect "the rule" is another of those more difficult in practice than in conceptualising (the last example breaks "if an element is the preceding element +"shire"; omit")

Would it be a "misuse" to use the standardised place name? (I suspect so)

You have also got a redundant "address field" - which could be populated where necessary with the required article - then report <[address] ><[place]>? Again "misuse of a field" that would not export well.

[tatewise]

The <[address] ><[place]> suggestion is a neat trick but has problems.

1. The < brackets > are not needed as FH will rationalise the spaces.
2. [address] and [place] should be {address} and {place} with curly brackets.
3. {place} produces in Wesleyan Chapel, ... so {address} {place} may produce the in Wesleyan Chapel, ...
4. Instead use in {address} {_place} which may produce in the Wesleyan Chapel, ...
5. the would have to be inserted in Address everywhere Wesleyan Chapel, ... was in Place.

[AdrianBruce]

...
Always make "Haslington Chapel, ..." the Place name, so sorting and type ahead works as normal.
In the Place record local Note enter "Article: the " to specify when the definitive article is needed.
In the Sentence Template, use {=GetLabelledText(%FACT.PLAC>NOTE2%,"Article:")} where the definitive article may appear. ...

That's rather interesting, thanks. I'd have to identify the possible sentences, but I can't imagine chapels of that name construction appearing in many places other than baptisms, marriages and burials.

It might also be useful for my military sentences where, as I think I discussed somewhere, the usage sometimes has "the" and sometimes not - e.g. "He served on HMS Victory" but "She served on the USS Enterprise". (Italics purely for emphasis). Although, as such values don't appear in any list, there's no particular need for consistency in the list in the first place (I enter them as values for a military service attribute).

[AdrianBruce]

...
What about "Tunstall Wesleyan Methodist Chapel, Tunstall, Staffordshire, England"?

That would indeed resolve the need for "the". I have two issue-ettes with the idea, though. Firstly, I'm not wild about the repeat of "Tunstall" (or whatever) for reasons of aesthetics and / or length. Perhaps to the point, I spent some time with Directories and other sources trying to find what the actual chapel names were - not a serious problem where it was just a village with one chapel per denomination, but more of an annoyance in a place like Crewe where each denomination might have several chapels (of course, most varieties of Methodist and Baptist were present, lengthening names still further). So I'm not anxious to supplement what I hope were the real names with some plausible but not actually accurate names.

...
"Durham Cathedral, Durham, County Durham, England"
or ...
"The Cathedral Church of Christ, Blessed Mary the Virgin and St Cuthbert of Durham, Durham, County Durham, England"?
I'd probably go with the former.

Oh, not the last? You do disappoint me! (Note smiley please!!!) (In my case, I have 3 different names for Manchester "Cathedral" that I need to remember to use at different dates! Not the same issue, of course.)

PS - re abbreviations. FYI - I'd quite like to have "... Wesleyan Methodist ..." in the places record and reports but abbreviate "Wesleyan Methodist" to "WM" in my charts.

[AdrianBruce]

I suspect that I shall investigate Mike's idea, thanks. It might also help me with things like "The Probate Registry" - though the conversion effort there might exceed the time saved by not needing to remind myself about the definite article.

[arthurk]

Proposal:
Place Records should be provided with an optional "sort order name". (I would enter "the Haslington Chapel..." as the Place Name, so it is used in narrative sentences and enter "Haslington Chapel..." as the sort name).

Then the Places tab should show the places in an order defined by the sort order name (if present) or the normal name. Also, when entering or updating places against facts (events, attributes, etc.) the type ahead facility should look at both the normal name and the sort order name, i.e. if I type "Haslington Cha..." it suggests "the Haslington Chapel..." and if I type "the Haslington Cha..." it also suggests "the Haslington Chapel...".

I may be missing something, but is a "sort order name" needed in Place Records? When I start typing a name in the filter box I find that it already brings up all places where any part/word begins with that string.

Is the real problem actually with the place picker that you use when adding/editing a fact, which looks only at the start of the place name? Then when you bring up the list using the ... button everything is sorted strictly alphabetically. You can click the column headers to sort according to each column, which might be useful if you format places strictly with a set number of parts (I don't), but whether you do that or not, would an alternative solution be for the place picker to be able to filter like the Place Records window does?

[tatewise]

PS - re abbreviations. FYI - I'd quite like to have "... Wesleyan Methodist ..." in the places record and reports but abbreviate "Wesleyan Methodist" to "WM" in my charts.

A similar technique can be made to put abbreviations into Diagram Text Scheme items for Charts.

In the Place record, add a labelled Note line for the abbreviation such as:
Abbr: WM, Tunstall, Staffs

Then in the Text Scheme Template use the expression:
=CombineText( , GetLabelledText( %INDI.BIRT.PLAC>NOTE2%, "Abbr:" ), , %INDI.BIRT.PLAC% )
which means display the Birth Place Note labelled Abbr: text if it exists, else display the Birth Place name in full.

[davidf]

What would you do with a baptism in the Cathedral at Durham?

"Durham Cathedral, Durham, County Durham, England"
...
or
"The Cathedral Church of Christ, Blessed Mary the Virgin and St Cuthbert of Durham, Durham, County Durham, England"?

I'd probably go with the former.

Oh, not the last? You do disappoint me! (Note smiley please!!!)

Of course not - the extra comma really messes things up! Otherwise ...

So I'm not anxious to supplement what I hope were the real names with some plausible but not actually accurate names.

[AdrianBruce]

...
I may be missing something, but is a "sort order name" needed in Place Records? When I start typing a name in the filter box I find that it already brings up all places where any part/word begins with that string.

Is the real problem actually with the place picker that you use when adding/editing a fact, which looks only at the start of the place name? ...

No, I think you're probably right - filtering the Place Records is, on reflection, probably perfectly adequate. As you say, the real issue is with the type ahead function, when I'd like to type "Haslington Cha..." and see it suggest "the Haslington Chapel..." and if I type "the Haslington Cha..." it also suggests "the Haslington Chapel...

But I'm unsure if a matching anywhere might not bring up too many. Imagine typing "Cheshire..." and getting all the partial matches to "Cheshire". Having what I had called a "Sort Order Name" as well as the ordinary name might suffice by providing just two options.

[arthurk]

But I'm unsure if a matching anywhere might not bring up too many. Imagine typing "Cheshire..." and getting all the partial matches to "Cheshire". Having what I had called a "Sort Order Name" as well as the ordinary name might suffice by providing just two options.

Except that in the example you're using, you'd presumably type "Haslington..." rather than "Cheshire..."

[tatewise]

But I'm unsure if a matching anywhere might not bring up too many. Imagine typing "Cheshire..." and getting all the partial matches to "Cheshire". Having what I had called a "Sort Order Name" as well as the ordinary name might suffice by providing just two options.

But it would not bring up too many multiple options. It would only show the single most recent match (just as it does now).
The more you type, the option shown changes until either you accept the type ahead or no options match.
i.e. When you type C for Cheshire, it will show the most recent Place name that contains a C or c anywhere, whereas at present it only shows one starting with C or c.
As you type more letters, h e s then the type ahead changes to the most recent match.

There could be a Tools > Preferences option to choose between leading characters and anywhere matching.

The concept of matching anywhere may need some refinement. Does it really mean matching anywhere or only at the start of words (which is what the Records Window Places tab Filter uses)?

[AdrianBruce]

...
The concept of matching anywhere may need some refinement. Does it really mean matching anywhere or only at the start of words (which is what the Records Window Places tab Filter uses)?

Food for thought - it must surely mean matching only at the start of words?

...
There could be a Tools > Preferences option to choose between leading characters and anywhere matching. ...

As a minimum, yes. But I suggest that might be a long winded approach and we would benefit from a switch closer to hand, perhaps linked to the "drop-down" somehow? Let me try and work through a couple of hopefully realistic cases here - as much for my benefit as anything.

Case 1a - Let's say we have matching only at the start of the place-name (as now), and that I want to enter a place-name of "Massachusetts, USA" against an event. It's already in my file as exactly that, so I type "Mas" and I will get "Massachusetts, USA". Seems good to me.

Case 1b - Let's say we have matching only at the start of the place-name (as now), and that I want to enter a place-name of "Massachusetts, USA" against an event, but it's not in my file as that. I type "Mas" and I won't get anything. I have to go to Google and find out how to spell "Massachusetts", even though I have a number of Massachusetts place-names in my place tab. (So I could go there instead of Google but I'm trying not to interrupt the data entry process).

Case 1c - Say we have matching at the start of any word in the place-name, and that I want to enter a place-name of "Massachusetts, USA" against an event, but it's not in my file as that. I type "Mas" and get one of my existing place-names in that state - probably "Fall River, Bristol Co., Massachusetts, USA", so I accept that suggestion and then delete the first two nodes, leaving me with "Massachusetts, USA" - so I didn't have to guess how to spell it.

Case 2a - Let's say we have matching only at the start of the place-name (as now), and that I want to enter a place-name of "Springfield, Hampden Co., Massachusetts, USA" against an event. I don't have that, so I'm probably going to type "Springf" and get "Springfield, , Ohio, USA" as a suggestion before I realise it's not in my files. How do I get the spelling of Massachusetts? Either Google or find something else in that state (such as "Massachusetts, USA"!) - all a bit of an interuption.

Case 2b - Say we have matching at the start of any word in the place-name, and that I want to enter a place-name of "Springfield, Hampden Co., Massachusetts, USA" against an event. I don't have that, and go through the process of Case 2a before I realise that I don't have it. I want to try to avoid mis-spelling Massachusetts, so type something like "Mas" and get my last Massachusetts place-name - probably "Fall River, Bristol Co., Massachusetts, USA", so I accept that suggestion and then modify the first two parts of the name.

Those cases are, I feel, realistic use cases for why I would like the ability to match at the start of any word in the place-name. But it seems to me that Case 2b (match at start of a word) is no different from Case 2a (match at start of a name) in that in both cases, I've typed "Mas" to pick up something with the correct spelling of the state's name. The only benefit would be in the undetailed Case 2c where we have matching at the start of any word, I want to enter a place-name of "Springfield, Hampden Co., Massachusetts, USA" and "Massachusetts, USA" isn't in my file. In that case, I'd get Fall River as a suggestion to edit.

Case 1c (match at start of a word) provides a benefit over case 1b (match at start of a name).

Match at the start of a word also provides a clear advantage in the case (Case 3 if you must) where I type "Wesleyan Chapel, Has" and get a suggestion of "the Wesleyan Chapel, Haslington, Cheshire, England". But if I follow the suggestion earlier in this thread of assigning "the" as a definite article in labelled text in the place-record, then that use case goes away...

So, frankly, I'm not convinced that place-name matching at the start of a word is that advantageous after all... Especially if I've already got US state names in my place list, even though I'm not necessarily using them at that level. (And I did just that for some US States or Canadian Provinces recently).

Having said that, if someone doesn't have the State or Province names, and / or doesn't use "the" as a definite article in labelled text in the place-record, then matching at the start of any word is going to be a benefit for them. Providing it's quick and easy to switch between the two match methods.

[tatewise]

I don't understand: "we would benefit from a switch closer to hand, perhaps linked to the "drop-down" somehow?"
What "drop-down"? I recall you've referred to "drop-down" before.
There is no "drop-down", only type ahead text in the Place box.

So, I'm getting the drift that the OP Subject: 'Sort order for Place Records' has been discarded, and the only benefit for matching words anywhere in Place names (instead of just the start) is where there might be a prefix such as the at the start, which is a rather small set of cases, or have I misunderstood.

[arthurk]

I've rarely needed to enter US states, and the difficult to spell ones even more rarely, so I don't feel I can offer much in relation to the workflow in the cases you identify. However, based on the way I (mostly) do things, does the following help for English villages, churches and other places of worship, in cases such as you mentioned originally:

When the church is an ancient parish church I usually don't bother to specify its dedication. Unless there's evidence to the contrary, marriages and burials took place there, and baptisms, even if performed elsewhere, will appear in that parish's registers.

However, where a parish has been divided I include the dedication in brackets, for both old and new churches. Thus, although for much of its history "Bradford, WRY" might have been a sufficient definition, I always refer to it as "Bradford (St Peter), WRY", and it takes its place in the list alongside "Bradford (St Augustine Undercliffe), WRY" and all the rest. This has the advantage of keeping all the Bradford churches together, with Bradford at the start of the name. (Note the lack of comma in the dedication: possibly there should be one, but including it would mess up the way FH identifies the parts of the name.)

The same principle applies to non-conformist churches, so I also have "Bradford (Otley Road Wesleyan Methodist), WRY" etc etc.

This works for me because the data is mainly for my eyes within FH or summary/outline reports. In narrative reports it might not flow very well, but it would mean you don't have to worry whether or not to include 'the'.

Just a thought...

[AdrianBruce]

...
What "drop-down"? I recall you've referred to "drop-down" before.
There is no "drop-down", only type ahead text in the Place box.
...

Yes, I'd got it in my mental picture that a drop-down (of one line!) appeared and can't drive the picture or the name away. As you say, it's just a suggestion that appears in the same box.

... the only benefit for matching words anywhere in Place names (instead of just the start) is where there might be a prefix such as the at the start, which is a rather small set of cases, or have I misunderstood.

Not quite -

... if someone doesn't have the State or Province names, ... then matching at the start of any word is going to be a benefit for them. Providing it's quick and easy to switch between the two match methods.

In other words, how do you spell Massachusetts? How can you get a prompt for that spelling out of FH's Place Records if you haven't already got an entry for "Massachusetts, USA"?

But as you suggest, the real impetus for me writing this was my use of Places that might - or might not - begin with "the", and I reckon your suggestion for the Labelled Text resolves that issue for me. (At the expense of some muttering from me about curly brackets, percentage signs, etc, etc.) So I have to say that I don't really see a lot of priority for this suggestion - not when compared to other issues.

[davidf]

This works for me because the data is mainly for my eyes within FH or summary/outline reports. In narrative reports it might not flow very well, but it would mean you don't have to worry whether or not to include 'the'.

I'm in the same camp; I don't let FH "write my family history".

Part of the issue is that we tend to write things in the opposite order to that in which we would want them to sort (increasing specificity?)

We write "John Smith" but sort Smith, John
We (in UK) write "25/12/2022" but sort 2022-12-25
We write "North Bailey, Durham City, County Durham, England" but sort England, County Durham, Durham City, North Bailey

Places have the added complication of missing elements which get marked by apparently surplus commas - none-the-less required if places are to "sort correctly":

Code: Select all

[Building], [Road/Street],       [Locality],         [Town],     [County],           [State], [(upper) Country]
Home Farm , The High Road,       Rowledge,           Farnham,    Surrey,             England, [missing "nation"]
Spires End, Ford Lane,           [missing locality], Farnham,    Surrey,             England, [missing "nation"]
Mar-a-Lago, 1100 S. Ocean Blvd., [missing locality], Palm Beach, [missing "county"], Florida, United States

Note how a "street number" further complicates matters - we don't really want 1100 S. Ocean Blvd sorting with other "1100 some-street" but with other addresses on "S. Ocean Blvd, Palm Springs" - makes spotting neighbours easier. It would not matter if you had an extra element for holding street numbers, but you then have to avoid the printing of the "formatting comma" between the street number and the street. (You could have a separate element as well for "definite articles" for chapels etc. - subject again to not printing the "formatting comma".)

We almost want a "place picker" which when selected presents a dialog with the elements arranged in order of increasing granularity (Country >>> Building) rather than the current (Building >>> Country). It could usefully (?) default to the last place used, and you tab through until you get to where you want a change, when you then overtype the default and it reverts to the current sort of "predictive" input.

So you might tab through:

Code: Select all

Nation:   [blank]
State:    [England]
County:   [Surrey]
Town:     [Farnham]
Locality: [Ro....]

And when you start typing Rowledge, it completes it - unless you type something else (Unlikely once you have typed the initial letters Ro - Farnham only has one "locality" starting "Ro" (it has "Runfold, and miss-spelt "wRecclesham"), and one you have the locality you want, you tab to the Street/Road - which has been auto-populated (either with the last address in Rowledge used, the most popular one, or just the first one), and you either over-type or tab on.

It would be a different experience to the current process but [tab] is no more difficult than [comma] [space]?

It would be useful if the dialog had user definable "groups of labels" ("Place templates" in FH7 parlance?) for "Nation", "State", "County" etc. You might even be able to select these from a predefined group of "Place Templates" so that when entering "English places", you got different labels from when entering "American places" or "(Current) Canadian Places" etc. That would enable consistent "comma'ing" of missing elements.

You wonder with such a dialog whether you could "enter" it in the middle at say "Town", type "Far .." until "Farnham" shows and then either cursor up to correct the county if necessary (and the country etc.) or tab down to complete.

Is this within the capabilities of a plug-in (some hefty processing and list reading/creating - and can it then enter data in the input field that had focus prior to running the plug-in)?

[AdrianBruce]

...
When the church is an ancient parish church I usually don't bother to specify its dedication.

Interesting - personally, I prefer to have the dedication there just for interest.

...
Thus, although for much of its history "Bradford, WRY" might have been a sufficient definition, I always refer to it as "Bradford (St Peter), WRY" ...

Yes, that would have been a good alternative for me. FamilySearch seem to write church names as "Bradford St Peter" which is a practice I dislike - I just ask about "Chalfont St. Giles"? (It's a village, not a church!) But your brackets solve that issue...

[tatewise]

Is this within the capabilities of a plug-in (some hefty processing and list reading/creating - and can it then enter data in the input field that had focus prior to running the plug-in)?

No, plugins in general cannot interact with the user dialogues.
(DEA can up to a point but not in this context.)

[arthurk]

...
When the church is an ancient parish church I usually don't bother to specify its dedication.

Interesting - personally, I prefer to have the dedication there just for interest.

If you were following my method with the brackets, there'd be nothing to stop you adding the dedication like that for any church.

[AdrianBruce]

...
If you were following my method with the brackets, there'd be nothing to stop you adding the dedication like that for any church.

That's true...

[Jean001]

Slightly off topic. This post is concerning the entering/sorting of placenames, not with the addition of a definite article to a placename in a sentence.

I read on FHUG quite some time ago that some FH users enter placenames backwards (i.e. Country, State, County, Town etc.)

I experimented with it, and I find that entering places this way is very quick. It works well with FH's 'type-ahead/predictive text. (I do all my data entry 'manually', i.e. without using AS.)

I am very well aware of the caveats so it not a practice that I recommend others to follow.

However, it occurs to me that 'Reverse Display Order' (which is available via Tools:Work with Data:Places) somehow might be incorporated into the 'Places Data Entry System' to ease the process.

[Vyger]

I fully understand the problem and frustration, this is a very emotive subject for me and I still manage my SITES, Geocoding, Media and Notes outside Family Historian. In that environment I can enter XXX Parish or XXX Town and get a predictive list of Addresses (Sites) relational to that Place entry as I type. I can also select one or several of those Places and obtain Mapping for all the Sites relational to those Places providing I have geocoded those Sites (Addresses).

In my opinion the SITE detail should not be part of the Place name but I understand for the purpose of geocoding this needs to be the case in the current build of FH. Competitor programs manage these sub divisions (Sites) as being relational to the Place which is the correct and logical model, I would believe this will become part of the FH Place handling at some stage in the future in order to level functionality across platforms.

FWIW, in my relational Site lists I enter the significant followed by the more specific usually in parentheses, some examples are below;

Main Street (#123) as opposed to 123 Main Street
Presbyterian Church (1st) as opposed to 1st Presbyterian Church
Methodist Church (New Connexion) as opposed to New Connexion Methodist Church

In the first example this groups the many hundreds of street addresses I have together helping me visualize the family groupings and associations as well as overcoming the sort fragmentation of street renumbering etc. In the second example I simply need to start typing the denomination to see the look ahead list of the Presbyterian Churches relational to the chosen Place or Parish, so all very intuitive and productive.

When I begin researching an area I start with Sites and enter the Parish with all Churches, Burial Sites etc as Place Details, I geocode these Sites and use the Notes to temporarily add specific research worthy details. As I complete and work these sites I update these details sometimes removing them to Research Logs.

FTR, my worldwide research of 110K people only results in a Place list of 3.5K but with many Address (Site) details relational to those Places. Therefore my Place List remains clean, easily understandable and not convoluted by superfluous Site details. I have watched presentations from long term FH users, one who stated he had 97K Places in his Place List, the mind boggles.

There is an Existing Wish List Item for Add Media and other data to Address fields, this is the key issue currently within FH and if addressed will resolve many of the present challenges. When addressed it will also present new challenges and development opportunities, improved online hints for those who currently combine Address and Place data to facilitate Geocoding, Notes and Media of Addresses, it will also create new research possibilities.

[AdrianBruce]

...
So, I'm getting the drift that the OP Subject: 'Sort order for Place Records' has been discarded, and the only benefit for matching words anywhere in Place names (instead of just the start) is where there might be a prefix such as the at the start, which is a rather small set of cases, or have I misunderstood.

I've just reminded myself from other threads why I originally wished for "matching words anywhere in Place names".

The market is bigger than place-names beginning "the" (which is still a small market). It's also place-names beginning with leading commas....

If I had entered ", , Massachusetts, USA" to get everything in the "right" column, then I'd like to be able to type "Mass" (say) and get ", , Massachusetts, USA" back as a suggestion. As it is, I'd need to type ", , Mass" and the problem with that is that I can never remember whether it's space-comma-space-comma-space at the front or comma-space-comma-space or ....

My inability to program my muscle memory over the exact combination of commas and spaces is one reason why I've never used leading commas. (I have other reasons by the way).