From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

Discuss the Shapefile page here:

File types and external links[edit]

I am going to clean up the file types used to make shapefiles to include both the types made in ArcVIew GIS 3.x and those made by using ArcGIS. They are different for the spatial and tabular indexes. user Rcc105

The external links were changed less into a listing of links as they should be and instead became a discussion. I also think it was done to promote a certain product. I am changing the external links back to the old version. --Ray 21:36, 7 February 2006 (UTC) yay!

uint32 should be int32?[edit]

Reading the ESIR shapfile spec, page 2 says "Integer: Signed 32-bit integer (4 bytes). This article, however, refers to unsigned integers throughout. This seems to be an error! MatthewVernon (talk) 11:58, 31 March 2008 (UTC)

Page title violates convention?[edit]

The page title (ESRI shapefiles) appears to violate the naming convention: prefer singular nouns. This seems even more likely in light of the first sentence in the article, which refers to shapefile in the singular:

  • The ESRI Shapefile is a data-type [...]

I suggest moving this article to ESRI shapefile and making ESRI shapefiles a redirect to it, unless the current article title should be plural for some reason I am not seeing. If the title should be plural, could someone give the reason here, in case someone else should wonder why the title appears to violate the convention. -- Teratornis 15:23, 18 May 2006 (UTC)

I agree, it should not be plural. Also, I've never called it an ESRI shapefile ... I only refer to it as simply a Shapefile, which is also the simplest lookup name. I really don't see why it is necessary to have ESRI as a prefix, since that is not common. Therefore, I suggest moving this article to Shapefile as there are no other things out there with that same name. I will do this in the comming months, unless I see any claim why the ESRI prefix should be kept. Mwtoews 22:26, 22 July 2006 (UTC)

I agree with changing it from ESRI Shapefiles to ESRI Shapefile. I do not concur with dropping ESRI. ESRI established the shapefile much in the way Microsoft established Excel files. Other products now use shapefiles and variants of the shapefile much in the way other non-MS office products use Excel spreadsheets. I think there should be a second and separate shapefile entry. --Ray 10:57, 23 July 2006 (UTC)

Okay, I'm somewhat agreeing with your comments to keep the ESRI prefix; however:

  1. ESRI developed and standardized the format, and certainly deserve credit, but it is not exactly a proprietary format exclusive for ESRI (this is counter to what the article says .. I will change this next)
    • Read this which states "The Shapefile spatial data format is open and published by ESRI ... ESRI will announce and publish changes as they are developed. ESRI encourages developers and users to create interchange capabilities to both read and create shapefiles."
      • Don't you just want to run up and give ESRI a big hug?
      • (Nothing like that has ever been stated from Microsoft about Microsoft Excel, which needs to be reverse-engineered by other third-party commercial and non-commercial products that want to "work" with MS Excel files.)
    • ESRI themselves rarely refer to the file as being an "ESRI Shapefile"
    • There are no trademarks related to "Shapefile" found through ESRI; look on Page 2 of the White Paper documentation provided by ESRI and you will not find any mention of "Shapefile" being a service mark of ESRI. I also don't think there are any patents on the format, and is an open spec; kind-of like PNG.
  2. The common name should be the the article name. Take NASA for instance, as it is the main article namespace (not the more complex National Aeronautics and Space Administration). It far easier for linking articles to use [[Shapefile]], rather than [[ESRI Shapefile|Shapefile]]
  3. The Shapefile namespace is open, and I doubt there will be any disambiguation on articles related on that term for quite a while; unless there is an upcoming movie called "Shapefile". If this were the case, then it would be ideal to disambiguate the article with an ESRI prefix.
  4. Simplicity is always better than complexity. (Or at least that's what I'd like to think.)

About splitting: no, I don't think this is necessary. I think that it is assumed that any Shapefile is an ESRI Shapefile, unless otherwise noted. I've never seen a hybrid shapefile format, nor should they exist. It would be foolish for any developer to drift away form the specifications given by ESRI, especially since the White Paper is very specific about data structures (no reverse engineering needed here). Again, unless I see a good argument about keeping the ESRI prefix, I'll move it at some point in the future to [[Shapefile]]. Mwtoews 21:11, 23 July 2006 (UTC)

I'd support a move to shapefile. -- nae'blis 20:02, 28 July 2006 (UTC)
I'd support either Shapefile (file format) or Shapefile, preferring the former. – Axman () 10:01, 30 July 2006 (UTC)
While I prefer the move to ESRI Shapefile, the tide seems to be to move to Shapefile. Is it time to do that or do we need to wait & discuss a bit more? Ray
The move must be done by an administrator because the Shapefile namespace is currently occupied (by a re-direct to ESRI shapefiles). I've already put in an application to Wikipedia:Requested moves, so it'll probably end up there in a few days or so. (Sorry Ray, I realize this is your baby article that you started a while ago). Mwtoews 16:41, 30 July 2006 (UTC)
YYou do not have to apologize. Any affection I had for Wikipedia went away with the administrator attack I went through with the University GIS page. It was not actually just admins, but a group of self appointed fools. I have given up hope that anything on Wikipedia will be without opinion and/or vulgarities. This was one of the few pages I watched anymore, not because I really cared but because I forgot to remove it from my Watchlists. The only thing I watch anymore are bike racers and bike races I know personally. --Ray 23:35, 30 July 2006 (UTC)

Wrong Move[edit]

I moved the article from ESRI shapefiles to ESRI shapefile, by accident. We'll have to wait for an admin to do the move to Shapefile. Sorry about that (although to be honest, ESRI shapefile should redirect to Shapefile anyway). -- tariqabjotu (joturner) 13:58, 2 August 2006 (UTC)

Done: Moved to Shapefile. Also, the following was copied from the replaced Talk page now deleted at Talk:Shapefile:
A user placed the remark shown below. He is correct however he fails to see that is stated on the ESRI shapefile page and fails to understand ESRI established the shapefile file format. --Ray 21:41, 7 February 2006 (UTC)
"Though Wiki describes Shapefiles very well under their title "ESRI Shapefiles", shapefiles are not specific to ESRI products."
Only coincidentally about the article title. —Centrxtalk • 02:26, 16 August 2006 (UTC)

Legal issues[edit]

Is the ESRI Shapefile format associated with any legal problems (underlying patents, copyrighted details, proprietary extensions) similar to the GIF or MP3 formats? Is the format compatible with GPL and the Wikipedia Foundation? For the detailed description of the file format, the current article refers to a technical specification, which I guess is copyrighted by ESRI. Is there any copylefted or public domain alternative? --LA2 14:20, 11 August 2006 (UTC)

Take a look at the ESRI Shapefile White Paper, in particular page 2. Here you will see where they own/have copyrights, trademarks, etc. ESRI has several, such as AML, and ARC GRID, but Shapefile does not appear here (i.e, it does not have a copyright or anything else). It is effectively an open standard, except that it is regulated by a profit organization. Since it is an open standard, there shouldn't be any legal problems whatsoever. I don't think this kind of thing needs to have a license, since it is just a data specification. — Mwtoews 02:40, 16 August 2006 (UTC)

Can I use the ESRI basedata that came on the CDs to make a map, and publish it to wiki? What copyright do I select? Broecher (talk) 16:50, 12 September 2009 (UTC)

Topology what?[edit]

The current article points out that Shapefiles don't store "topological information", but that word is neither linked nor explained. I'd like to make it a link, but to what? There is an article topological map. In the page topology (disambiguation), one can read that "Topology, is the study or science of places with applications in earth science, geography, human geography, and geomorphology" and also that "Topology, a term used in architecture to describe spatial effects which cannot be described by topography, i.e., social, economical, spatial or phenomenological interactions". But does any of this apply here? --LA2 15:02, 11 August 2006 (UTC)

  • I had the same reaction - What do they mean by topological information? On browsing I found that the article Geospatial_topology seems to describe what they are talking about -- basically, connectedness of features, independent of scale or units. Added links accordingly. Huttarl (talk) 21:54, 23 September 2009 (UTC)

ambiguous AutoCAD shape font file[edit]

AutoCAD's shape font source file (*.shp) and compiled shape font file (*.shx) are totaly different from ESRI Shapefile. It is confusing. Could you mention this in the article in good English? Viking rollo 14:00, 9 February 2007 (UTC)

I don't know anything of these formats; however, any article should never be about what the article is not about (although I see your confusion, since AutoCAD and ESRI are similar companies with similar products; also there are dxf <-> shp converters too). I don't think Wikipedia has anything that you speak of, so if you feel ambitious, you could start an AutoCAD font article.+mwtoews 21:06, 9 February 2007 (UTC)


(.shp-file... ) however this file alone is incomplete for distribution, as it depends on the other supporting files. - the .shx-file however does not depend on others, neither on the .shx (index file depends on the shp, not v.v.) nor on the .dbf (extra attribute information). Cuold any native english speaker/writer correct this? 11:35, 12 April 2007 (UTC)

It's probably simplest to remove the dependency term from that statement, which I've done, and just state that the supporting files are required. 17:34, 12 April 2007 (UTC)

Move article[edit]

  • Keep at Shapefile—it is more common. It is true that "ESRI Shapefile" is an official name, however, ESRI usually refers to the format as simply a "shapefile"—the former only appears on the cover of this white paper, while the bulk volume refers to "shapefile" only. Most of ESRI's other documents also use "shapefile" far more than the longer title (e.g., this doesn't even mention ESRI for the format, and it is the current documentation from ESRI). Also, all the interwikis use "Shapefile" for their name. I think Shapefile (which was the former name of the article—now reverted back) is the better, since it is most common. Prefixing the company name in front of the format name sounds as funny as Adobe PDF. Furthermore, there are many software products that use the format, that have no affiliation with ESRI. It's an open standard. +mt 16:08, 2 November 2007 (UTC)
(Note: I reverted the prior move from Shapefile to ESRI Shape format, since this is a really incorrect name. This discussion is whether to keep this article at Shapefile or to move it to ESRI Shapefile.)+mt 18:49, 5 November 2007 (UTC)

It was requested that this article be renamed but there was no consensus for it to be moved. Angus McLellan (Talk) 14:42, 12 November 2007 (UTC)

Remove directions for .sbn format regeneration?[edit]

I would vote to remove the vendor/software specific directions from the "Shapefile spatial index format (.sbn)" — it seems inappropriate to have directions for a particular software product in the entry about a file format.

--JeremyCole (talk) 07:00, 28 November 2007 (UTC)

Agree, delete. It shouldn't be there. I'm sure if it is useful, it could be found on a forum or other place. I suppose the most you could say that the .sbn file is undocumented, and specific only to ESRI products (and I guess not required to define shapefiles).+mt 07:30, 28 November 2007 (UTC)

Which applications? How old? How common?[edit]

"Each individual file should conform to the MS DOS 8.3 filename convention (8 character filename prefix, fullstop, 3 character filename suffix such as shapefil.shp) in order to be compatible with past applications that handle shapefiles"

This COULD be recommended for ALL modern software whose files might, under some conceivable circumstance, be read by a program still limited to 8.3 filenames, but generally isn't. Why here? Who uses what that justifies this caveat? —Preceding unsigned comment added by (talk) 20:53, 6 March 2010 (UTC)

The use of dBase file requires the 8+3 naming convention. For example, Microsoft's JET provider won't open a dBase file with a longer file name. Even the newest versions have this limitation. However, if you use a more intelligent implementation (GIS software, or LibO Calc) you don't have to worry about long file names. ITinerisKft (talk) 09:47, 12 September 2011 (UTC)
The other issue is that you almost never work with single shapefiles -- you almost always work with collections, and the application software you're using typically constructs the names for the several files making up the collection (whatever.shp, whatever.shx, whatever.dbf, etc.). So, unlike an application that simply opens the one file whose pathname you feed it, shapefile-manipulating applications also have to manipulate and construct shapefile-related pathnames, and it's likely that some of them were written in such a way that they still depend on (that is, enforce) the old 8.3 restriction. —Steve Summit (talk) 18:05, 11 September 2015 (UTC)

Does not specify format(s) for storage of the shapes[edit]

I'm assuming it's doubles and pairs of them for 2d shapes and triplets for 3d (Edit: Checked the whitepaper: A little more complex than that) but it would be nice if it was actually written in there explicitly, especially as endianness varies. (talk) 19:23, 1 May 2015 (UTC)

Mixing shape types[edit]

Our article correctly states that "Because the shape type precedes each geometry record, a shapefile is physically capable of storing a mixture of different shape types", but that the official specification prohibits such mixing. It's inconceivable to me that no one has taken steps (if only informally) towards relaxing this arbitrary restriction. If there are any applications which support "mixed" shapefiles, any established conventions for doing so, it would be great to document them here. —Steve Summit (talk) 17:57, 11 September 2015 (UTC)


The article mentions that M can be optional, but it is not entirely clear how. This is by supplying a magic value that indicates unknown for points and header MBRs. Allocating insufficient size for the additional information appears to be the way of doing this on a per row basis which is odd for a format that requires all entries to be of the same type. — Preceding unsigned comment added by (talk) 22:52, 26 June 2017 (UTC)

Only M is optional, and is read if the record length (or "content length") is long enough to read the values. See page 5 of the white paper, which is the primary reference for the format used in this encyclopedia article. +mt 03:02, 27 June 2017 (UTC)