GeoNames Forum

Hello,
To separate my comments, I have created this topic.
So, I have noted some duplicated entries in the french postal codes.
In the following list, I have highlighted the entries which must be deleted.

(Apparently, some city are considered as belonging to 2 departments/states (admin name/code 2). However, a city cannot belong to 2 states... The first two digits of the postal code determine the home department/state of the city)

FR 24140 Jaure Aquitaine 97 Dordogne 24 44.9 0.5 5
FR 24140 Jaure Aquitaine 97 Dordogne 24 Arrondissement de Périgueux 243 45.0667 0.5667 5

FR 63340 Collanges Auvergne 98 Puy-de-Dôme 63 Arrondissement d'Issoire 633 45.4 3.7833 5
FR 63340 Collanges Auvergne 98 Puy-de-Dôme 63 45.4333 3.2333 5

FR 56160 Locmalo Bretagne A2 Morbihan 56 48.0667 -3.1667 5
FR 56160 Locmalo Bretagne A2 Morbihan 56 Arrondissement de Pontivy 562 47.7 -3.3514 5

FR 18190 Chambon Centre A3 Cher 18 46.9667 2.5167 5
FR 18190 Chambon Centre A3 Cher 18 Arrondissement de Saint-Amand-Montrond 182 46.7833 2.3333 5

FR 20290 Vignale Corse A5 Haute-Corse 2B Arrondissement de Bastia 2B2 42.5333 9.3833 5
FR 20290 Vignale Corse A5 Haute-Corse 2B 42.4667 9.3333 5

FR 27160 Breteuil Haute-Normandie A7 Eure 27 Arrondissement d'Évreux 273 48.8367 0.9143 5
FR 27160 Breteuil Haute-Normandie A7 Eure 27 48.8333 0.9167 5

FR 76450 Bertheauville Haute-Normandie A7 Seine-Maritime 76 49.75 0.5833 5
FR 76450 Bertheauville Haute-Normandie A7 Seine-Maritime 76 Arrondissement de Dieppe 761 49.85 0.65 5

FR 76500 La Londe Haute-Normandie A7 Seine-Maritime 76 49.6313 0.9579 5
FR 76500 La Londe Haute-Normandie A7 Seine-Maritime 76 Arrondissement de Rouen 763 49.3 0.9667 5

FR 76500 Orival Haute-Normandie A7 Seine-Maritime 76 49.7333 1.2 5
FR 76500 Orival Haute-Normandie A7 Seine-Maritime 76 Arrondissement de Rouen 763 49.3 0.9833 5

FR 54740 Saint-Remimont Lorraine B2 Meurthe-et-Moselle 54 Arrondissement de Nancy 543 48.5 6.25 5
FR 54740 Saint-Remimont Lorraine B2 Vosges 88 Arrondissement de Neufchâteau 882 48.2455 5.8997 5

FR 88500 Blémerey Lorraine B2 Meurthe-et-Moselle 54 Arrondissement de Lunéville 542 48.5833 6.7333 5
FR 88500 Blémerey Lorraine B2 Vosges 88 Arrondissement de Neufchâteau 882 48.3667 6.05 5

FR 46100 Viazac Midi-Pyrénées B3 Lot 46 44.6333 2.1 5
FR 46100 Viazac Midi-Pyrénées B3 Lot 46 Arrondissement de Figeac 462 44.6667 1.8333 5

FR 81220 Prades Midi-Pyrénées B3 Tarn 81 43.6 1.9833 5
FR 81220 Prades Midi-Pyrénées B3 Tarn 81 Arrondissement de Castres 812 43.7 2.55 5

FR 44540 Le Pin Pays de la Loire B5 Loire-Atlantique 44 Arrondissement d'Ancenis 444 47.35 -2.1667 5
FR 44540 Le Pin Pays de la Loire B5 Loire-Atlantique 44 Arrondissement d'Ancenis 444 47.5893 -1.153 5

FR 02870 Crépy Picardie B6 Aisne 02 49.6 3.5167 5
FR 02870 Crépy Picardie B6 Aisne 02 Arrondissement de Laon 022 49.6044 3.5144 5

FR 06420 Valdeblore Provence-Alpes-Côte d'Azur B8 Alpes-Maritimes 06 Arrondissement de Nice 062 44.0712 7.172 5
FR 06420 Valdeblore Provence-Alpes-Côte d'Azur B8 Alpes-Maritimes 06 44.0667 7.2 5

FR 42600 Savigneux Rhône-Alpes B9 Loire 42 45.6167 4.0833 5
FR 42600 Savigneux Rhône-Alpes B9 Loire 42 Arrondissement de Montbrison 421 45.6 4.4167 5

FR 73660 La Chapelle Rhône-Alpes B9 Savoie 73 Arrondissement de Saint-Jean-de-Maurienne 733 45.4167 6.2833 5
FR 73660 La Chapelle Rhône-Alpes B9 Savoie 73 45.5167 6.0833 5

FR 73670 Saint-Pierre-d'Entremont Rhône-Alpes B9 Isère 38 Arrondissement de Grenoble 381 45.4146 5.8552 5
FR 73670 Saint-Pierre-d'Entremont Rhône-Alpes B9 Savoie 73 Arrondissement de Chambéry 732 45.417 5.8516 5

FR 62760 Thièvres Nord-Pas-de-Calais B4 Pas-de-Calais 62 Arrondissement d'Arras 621 50.1333 2.45 5
FR 62760 Thièvres Picardie B6 Somme 80 Arrondissement d'Amiens 802 50.1283 2.4567 5

The datas in bold must be deleted : For each duplication, I have chosen the entry which have the least informations, or which have the best information : For example, for the city "Le Pin", the entry which must be deleted has wrong latitude and longitude)

The datas in bold and red must be deleted too and correspond to my remark below.

In advance, thanks for your answer.

Marc, are you not agree with my analyze ?

I agree that those are duplicates and I have started to fix it:
http://www.geonames.org/recent-changes/country/fr/

For France the postal codes are extracted (by running a small program) from the alternatenames of the main toponym tables. The duplicates can either be caused by a duplication of a place in the main database (solution>delete the place) or by a valid second place with the same name. In the later case I am not yet sure what it means. What is the real postal code of these places?
Example:
http://www.geonames.org/3024337/collanges.html
http://www.geonames.org/3024338/collanges.html

Best Regards

Marc

OK.
Again, France appears to be a complex country.

If you have problem, I can help you because I'm french (So I'm complicated too

)

I have made research and I have a response.
There is indeed 2 "Collanges" : one of this collanges is a town. The other is a "lieu-dit" (in french). I think the translation of "lieu-dit" could be "hamlet" in english.
So, an hamlet is too small to be considered as a town, and the the French postal service associate no postal code with an hamlet.
(The exception is when a town become an hamlet. In this case, the french postal service keep the associated postal code)

To conclude for collanges case...
This entry correspond to the town :
FR 63340 Collanges Auvergne 98 Puy-de-Dôme 63 Arrondissement d'Issoire 633 45.4 3.7833 5

This entry correspond to the hamlet and must be deleted from the french postal code repository :
FR 63340 Collanges Auvergne 98 Puy-de-Dôme 63 45.4333 3.2333 5

(You will find a proof by vewing this (french) page : http://fr.wikipedia.org/wiki/Arlanc / The Collange which must be deleted, is a "lieu-dit" of the town called Arlanc)

If you have other cases like this, fell free to tell me.... I will investigate

Best regards.

Hello Marc,
Given the complexity of the work, I have decided to analyse the duplication that I have Highlighted.

SUMMARY
- Some of these entries are really duplicated. SO, one must be deleted
- Like COLLANGES, some place (hamlet, port, castle...) have the same name as a town... But it's incorrect to say that these places have a postal code. So we must deleted the corresponding entry
(I don't know how works your extractor. Maybe, your little program considers these place as city, but it's false)
- Some of this entries are simply false and must be deleted
- 2 entries are exceptions : The case of a city which belong to 2 states and which have only one postal code.

Be careful, for some correction, I have modified the geographic coordinates, or add admin information or fusion the duplicated entries informations. It will be good if you can report these modifications)

- JAURE
For me there is an error... There is only one town called Jaure in France.
No geographic coordinates in the french postal code repository matches the real geographic coordinates.
I have no explanation about the presence of the two JAURE in the repository (surely the same case of COLLANGES)
(Source : http://fr.wikipedia.org/wiki/Jaure)

The only entry for this town must be :
FR 24140 Jaure Aquitaine 97 Dordogne 24 Arrondissement de Périgueux 243 45.0575 0.5547 5

- COLLANGES
See my explanation in my previous answer.

The only entry for this town must be :
FR 63340 Collanges Auvergne 98 Puy-de-Dôme 63 Arrondissement d'Issoire 633 45.4 3.7833 5

- LOCMALO
I was wrong in my previous answer. I have selected the wrong entry to delete.
So, it must be only one town town called Locmalo. (The other "Locmalo" is not a town. It's the name of a port in the city called "Lorient")
(Source : http://fr.wikipedia.org/wiki/Locmalo )

The only entry for this town must be (I have change the geographic coordinates to make it more accurate and add other informations) :
FR 56160 Locmalo Bretagne A2 Morbihan 56 Arrondissement de Pontivy 562 48.0733 -3.1853 5

- CHAMBON
We are in the same case of COLLANGES.
(Source : http://fr.wikipedia.org/wiki/Chambon_%28Cher%29 )

The only entry for this town must be (I have change the geographic coordinates to make it more accurate) :
FR 18190 Chambon Centre A3 Cher 18 Arrondissement de Saint-Amand-Montrond 182 46.7878 2.3255 5

- VIGNALE
We are in the same case of COLLANGES.
(Source : http://fr.wikipedia.org/wiki/Vignale_%28commune%29 / http://fr.wikipedia.org/wiki/Bisinchi)
The wrong entry is in fact, the name of a castle near the town of BISINCHI.

The only entry for this town must be (I have change the geographic coordinates to make it more accurate) :
FR 20290 Vignale Corse A5 Haute-Corse 2B Arrondissement de Bastia 2B2 42.5389 9.3822 5

- BRETEUIL
The 2 entries are really duplicate. (the geographic coordinatess point to the same place)
I propose to conserve only this entry (which has more information) :
FR 27160 Breteuil Haute-Normandie A7 Eure 27 Arrondissement d'Évreux 273 48.8367 0.9143 5

- LA LONDE
We are in the same case of COLLANGES.
(Source : http://fr.wikipedia.org/wiki/La_Londe)
The wrong entry is in fact, the name of a castle near the town of CALLEVILLE.

The only entry for this town must be (I have change the geographic coordinates to make it more accurate) :
FR 76500 La Londe Haute-Normandie A7 Seine-Maritime 76 Arrondissement de Rouen 763 49.3067 0.9539 5

- ORIVAL
We are in the same case of COLLANGES.
(Source : http://fr.wikipedia.org/wiki/Orival_%28Seine-Maritime%29 / http://fr.wikipedia.org/wiki/Saint-Hellier)
The wrong entry is in fact, an hamlet of the town of SAINT-HELLIER.

The only entry for this town must be (I have change the geographic coordinates to make it more accurate) :
FR 76500 Orival Haute-Normandie A7 Seine-Maritime 76 Arrondissement de Rouen 763 49.3128, 0.9975 5

- SAINT-REMIMONT
There is an important error for this town.
In france, there is 2 towns called SAINT-REMIMONT. One of these towns is in the state "Vosges" (code 8

and the other in the state "Meurthe-et-Moselle" (code 54).
However, in the repository, there is 3 towns called SAINT-REMIMONT and among them, there is 2 towns in the state "Vosges".
One of these 2 towns has the first two digits of its postal code equals to 88 and the other 54.
So, the town "SAINT-REMIMONT" which is in the state "Vosges" and which has the first two digits of its postal code equals to 54 must be deleted.
(Source : http://fr.wikipedia.org/wiki/Saint-Remimont_%28Meurthe-et-Moselle%29 / http://fr.wikipedia.org/wiki/Saint-Remimont_%28Vosges%29)

The only entries for this town must be (I have change the geographic coordinates to make it more accurate) :
FR 54740 Saint-Remimont Lorraine B2 Meurthe-et-Moselle 54 Arrondissement de Nancy 543 48.4989 6.2505 5
FR 88800 Saint-Remimont Lorraine B2 Vosges 88 Arrondissement de Neufchâteau 882 48.2452 5.8997 5

- BLEMEREY
Like the town previously, there is an important error for this town.
The problem is exactly the same.
The town "BLEMEREY" which is in the state "Vosges" and which has the first two digits of its postal code equals to 54 must be deleted.
Source : http://fr.wikipedia.org/wiki/Bl%C3%A9merey_(Meurthe-et-Moselle) / http://fr.wikipedia.org/wiki/Bl%C3%A9merey_(Vosges))

The only entries for this town must be (I have change the geographic coordinates to make it more accurate) :
FR 54450 Blémerey Lorraine B2 Meurthe-et-Moselle 54 Arrondissement de Lunéville 542 48.5839 6.735 5
FR 88500 Blémerey Lorraine B2 Vosges 88 Arrondissement de Neufchâteau 882 48.3605, 6.0489 5[/i][/b]

- VIAZAC
We are in the same case of COLLANGES.
(Source : http://fr.wikipedia.org/wiki/Viazac)
The wrong entry must be an hamlet of the town of LIVERNON.

The only entry for this town must be (I have change the geographic coordinates to make it more accurate) :
FR 46100 Viazac Midi-Pyrénées B3 Lot 46 Arrondissement de Figeac 462 44.6422 2.0758 5

- PRADES
I was wrong in my previous answer. I have selected the wrong entry to delete.
We are in the same case of COLLANGES.
(Source : http://fr.wikipedia.org/wiki/Prades_(Tarn))
The wrong entry is an hamlet of the town of ESPERAUSSES.

The only entry for this town must be (I have change the geographic coordinates to make it more accurate) :
FR 81220 Prades Midi-Pyrénées B3 Tarn 81 Arrondissement de Castres 812 43.6142 1.9744 5

- LE PIN
We are in the same case of COLLANGES.
(Source : http://fr.wikipedia.org/wiki/Le_Pin_(Loire-Atlantique))
The wrong entry is a quarter of SAINT-MALO-DE-GUERSAC.
(Source : http://fr.wikipedia.org/wiki/Le_Pin_(Loire-Atlantique) / http://fr.wikipedia.org/wiki/Saint-Malo-de-Guersac)

The only entry for this town must be :
FR 44540 Le Pin Pays de la Loire B5 Loire-Atlantique 44 Arrondissement d'Ancenis 444 47.5893 -1.153 5

- CREPY
The 2 entries are really duplicate. (the geographic coordinatess point to the same place)
I propose to conserve only this entry (which has more information) :
FR 02870 Crépy Picardie B6 Aisne 02 Arrondissement de Laon 022 49.6044 3.5144 5

- VALDEBLORE
The 2 entries are really duplicate. (the geographic coordinatess point to the same place)
I propose to conserve only this entry which has more information) :
FR 06420 Valdeblore Provence-Alpes-Côte d'Azur B8 Alpes-Maritimes 06 Arrondissement de Nice 062 44.0712 7.172 5

- SAVIGNEUX
We are in the same case of COLLANGES.
(Source : http://fr.wikipedia.org/wiki/Savigneux_(Loire))
The wrong entry must be an hamlet of the town of CHEVRIERES.

The only entry for this town must be (I have added admin 3 informations) :
FR 42600 Savigneux Rhône-Alpes B9 Loire 42 Arrondissement de Montbrison 421 45.6167 4.0833 5

- LA CHAPELLE
We are in the same case of COLLANGES.
(Source : http://fr.wikipedia.org/wiki/La_Chapelle_(Savoie))
The wrong entry must be an hamlet of the town of CRUET.

The only entry for this town must be :
FR 73660 La Chapelle Rhône-Alpes B9 Savoie 73 Arrondissement de Saint-Jean-de-Maurienne 733 45.4167 6.2833 5

- SAINT-PIERRE-D'ENTREMONT
According to Wikipedia, this town belongs to 2 differents states.
Normally, the couple (town, postal code) must be unique, but this town appears to be an exception...
In France, cities are identified by a unique code called "INSEE code". And this town is considered as 2 different town (because there is 2 different INSEE code) but which have the same postal code.
(Source : http://fr.wikipedia.org/wiki/Saint-Pierre-d%27Entremont_(Is%C3%A8re) / http://fr.wikipedia.org/wiki/Saint-Pierre-d%27Entremont_(Savoie) )

So, we must delete no entry.

- THIEVRES
We are in the same case of SAINT-PIERRE-D'ENTREMONT
(Source : http://fr.wikipedia.org/wiki/Thi%C3%A8vres_(Pas-de-Calais) / http://fr.wikipedia.org/wiki/Thi%C3%A8vres_(Somme) )

I hope thanks to my analyse you will be able to modify the repository quickly and easily.
Best regards.

Hi

You can make the changes directly yourself:
http://www.geonames.org/manual.html

I had already fixed the first three.
In order to update the postal code export I will have to run a script that extracts them from the toponym table as soon as the changes are done. This is only for France and the dependent countries. The other postal codes are not yet closely linked to the toponym table and need to be edited separately.

Best

Marc

OK,
The only changes I can make, are modifying the geographic coordinates and admin information. I can't modify or delete a postal code. Do I understand?

As I really don't understand what is your postal code export process, will you be able to manage the deletion of the place which must no appear in the french postal code repository ?

If you are interested, I can give you the postal code (and the geographic coordinates) for French Polynesia and New Caledonie. I have made them by hand.

I can give you too an official french postal code repository. (But in private because it'sa not a free file)

I think I could make indeed the modifications...
I hope I have made good job.
I have deleted the postal code in the alternate name when it was wrong... I hope it will resolve the problem.

However, I couldn't do this actions :

- I can't delete the real duplicated places (Because I have not the right) :
BRETEUIL -> GeoNameId 3030275
CREPY -> GeoNameId : 3022570
VALDEBLORE -> GeoNameId : 2980931

- I can't move this town because it's locked....
Blémerey -> GeoNameId : 3032280
I would like to move to the point : 48.3605, 6.0489

Could you make this modifications for me ?

PS : Sorry for my ignorance. I just started to view the power of geoname tool.

Mark,
Would it be possible to know the feeding process of the geonames database and the export process ?
How does taking into account the manual data entry?

My question is linked to the fact that I don't see my recent modification in the last postal code repository.

Is there a document or webpage which can explain me this operation ?

In advance, thanks for your answer.
Best regards.

I have updated the database with the french postal codes. It needs to be started manually, it is not yet automated. The export is automated and runs daily, but the process from from the toponyms to the postal code database is manual and only for France. All other countries are updated directly on the postal code database (no edit interface, only batch updates)

Marc

OK.
For your information, I started updating postal codes from the official french postal code repository that I got.
It will take considerable time, but It takes a long time but it's worth.

Periodically, I will show you the changes I could not do so you can make them for me.

I will see after if I can record the postal code for French Polynesia and New Caledonie.

thanks. I will periodically run the processing when you have made some changes.

Best

Marc

@steptom :

French postal codes have been updated a few weeks ago according to laposte official list.

Please do not delete CEDEX postal codes such as 02023 LAON
because some geonames users need this full list to validate input forms.

@geotree
My update are based on the laposte official list.
Thanks to my activity, I receive this list monthly. This list is called "HEXAPOSTE".

In the last version of this list, "02023 LAON" is a CEDEX postal code which is not yet assigned to a "client".
It's the same for :
02021 LAON
02022 LAON
02024 LAON

Be sure that I would no delete these postal codes if I was not sure.
So are you ok with that ? Can I keep doing this?

If you want, I can send you the official list file for proof.

Please, I need an answer because I have detected other CEDEX postal code which are no longer available now...

Note : If these CEDEX postal codes must not be deleted, I think we should add the other CEDEX postal codes which are not assigned. (there is a lot of code in this case)

Hello Marc,
since I posted my last message, I have received no answer.
I have sent also a private message to "geotre" but I have received no answer too.

How can I have an answer to my problem because I'd like to continue my work ?

In advance, thanks for your answer.
I wish you a Happy New Year 2011.
Best regards

Hi Tom

I am sure Christophe will answer as soon he has time, I guess he went away over the holidays and had no internet access.
I don't know much about the French CEDEX, so we have to wait for Christophe to answer this question.

My best wishes for a Happy 2011. Thanks a lot for your contributions.

Best

Marc

Thanks for your answer Marc,
I'm reassured now.

In short time, I will send you a list of action that I could not do because of my access level.

I have a question about belgium postal code, maybe can you answer me...
In the belgium postal code repository, there is some city with a question mark in their name.
For example :
- Hasselt ?, Stokrooi
- Lanaken ?, Kesselt
- Flémalle ?, Les Cahottes
- Evergem ?, Doornzele
- Evergem ?, Belzele
- ...etc.

It's ugly to display that and, in addition, we don't see these names on the google map. Example :
- http://www.geonames.org/postalcode-search.html?q=Evergem+%3F%2C+Doornzele&country=
- http://www.geonames.org/maps/google_51.146_3.717.html

So, do you know why there is this question mark ? Is there a solution to delete them and to adjust the name of these cities in the postal code repository with the name displayed on the google map ?
Thanks for your answer.

Thanks. I have updated the DB and removed the questionmarks. I don't know where they are from. Probably an encoding issue with an input file.

Marc

The Belgium postal codes have not yet been fully linked with the toponyms in the main database (as they have for France, thanks to the hard work done by Christophe). You are welcome if you want to help with this. I do have some automatically generated links to get you started.

Marc