Subscribe to this thread
Home - General / All posts - connect to postgis or postgresql database with M8

843 post(s)
#03-Feb-19 04:38

I am struggling with connecting to a postgis database with Manifold 8.

Can anyone help me out with details of how to do so.

The helpfile does not seem to have details of what should go in the various fields.

I am just testing with a local database at the moment, so I am putting

localhost for the server. is this correct or do I need to include the port or something as well?

user and password I think I have figured out.

Database I assume is the name of the database on the server.

Filling this out with the same details that work in QGIS just gets me

"can't establish connection to datasource" in Manifold.

Are there some key details or steps I am missing? Are there limits on what versions of PostGres Manifold supports?

I can see from the forums that other people are having no problems with this, I suspect I have just missed something.

There is one webpage suggesting you need to add the postgres binaries location to the path variable for it to work. I have done this, but find it somewhat baffling as I would have thought Manifold should be able to connect to a server without having the binaries installed locally?

I do have a manifold server console installation on an old version postgresql that works, but needs to connect via odbc drivers.


843 post(s)
#03-Feb-19 05:11

Answering my own question to some degree.

I just tried using the ODBC drivers, and was able to connect.

The help file seems to suggest this is optional, is it actually the normal method most people use?


843 post(s)
#03-Feb-19 05:26

Next question - although I am able to connect and link tables with the postgresql odbc driver, manifold does not seem to recognise the geometry data, and just brings in the table. Is there another step to get the geometry working too?


843 post(s)
#08-Feb-19 07:00

Noone able to help with this?


5,491 post(s)
#08-Feb-19 07:24

I'd guess that most people who do work with spatial DBMS have moved on to 9.

Here's an idea: download Viewer and use that to connect to PostgreSQL. If there are any issues, launch a new thread to discuss that.


843 post(s)
#10-Feb-19 06:21

Is 9 considered finished yet? I haven't been keeping up with it closely, but was under the impression the layout and formatting was still not up to 8 standard, and 90% of our work is pretty basic presentation of data and maps rather than number crunching.


6,335 post(s)
#10-Feb-19 11:27

I think label and vector formating is way better in Mfd than in Mfd8 or ARcGIS now, features, display and controls. And I think - curved elements and 3D still to come - development has reached some maturity! Can't say much about raster data because remote sensing is over my head and I didn't have the chance to play with surfaces.

Anyway it's time to add sophisticated layouts and legends.


5,491 post(s)
#10-Feb-19 12:14

Is 9 considered finished yet?

9 will never be finished, in the sense that living products are never finished. If it's a living product there is always something more you want to add, right up to the point where the only way to add significant new things is to do the next generation, 10.

Whether you think 9 is finished enough for what you do depends entirely on what it is you do. People who use GIS to slice and dice data using SQL, grabbing data from DBMS, manipulating it and then putting it back in new, improved form thought 9 was "finished" enough for them over a year ago.

The maps I do for personal use tend to have very simple formatting, so for me 9 was plenty finished enough to replace 8 almost a year ago as well. Effects like this antique swiss look have been possible for a year. The new formatting capabilities that have come out in the last 60 days or so have staggeringly many possibilities, which is good. They certainly exceed 8. All of the other things in 9 have led me, hands down, to shift to 9. I like 8 and I'm proud of my role in 8 but after getting my head around 9 I find it way easier to use 9.

9 capabilities now far exceed what is in 8, but they do so in a different, more fundamental, more data centric direction. That's great for people who want that, but for people who miss some of the Photoshop-like features in 8 and similar, 9 is not yet there for them. Doesn't matter how many features 9 has if the feature you must have isn't one of them. But don't assume that if at first glance there is not a 1:1 analog to 8 that the capability you need is not in there.

What you'll find with 9 is there are often different ways of doing something than in 8. Until you learn 9, you might think 9 doesn't have those analogous capabilities. But now, more often than not, it does, and after learning how it is done in 9 it's pretty rare to find something that done the 9 way doesn't open up many possibilities not available in 8.

We are on the cusp of a push through layouts and legends (as discussed in other threads, that's what ideas of grouping, be that grouping of layers or grouping of layout elements, is about), and then we need a cycle through geometry and a big cycle through rasters. Along the way there will be the usual sprinkling of well-known, desired things, like georeferencing images/drawings, an IMS server and such. My guess is that by this summer there will be very, very few people working with 8 who could not better be working with 9.

After that come endless things cosmically beyond 8. Distributed processing, automatically, on multiple machines in your local network or in the cloud, all kinds of 3D stuff, specially fast and powerful LiDAR tools, very much more sophisticated vector and raster editing (no reason to need to use either Illustrator or Photoshop), AI, some very nifty ideas for DBMS, endless exploitation of web resources, systematic working down of all ESRI tools to provide Manifold analogs, numerous new GUI features, and on and on. What's nice about 9 is all that is very easily supported within the engine as it is, thanks to all that effort (years) that went into Radian.


843 post(s)
#11-Feb-19 01:38

Sounds good, I'll have another look when the layouts and legends are going.

Mike Pelletier

1,602 post(s)
#12-Feb-19 15:46

Glad you mentioned georeferencing images and drawings. Given 9's speed it would be great if it worked with a preview so that as each control point is added the result can be instantly seen. One could then drag a control point a tad to see if it improves the result or just turn off the control point and see the result. This would be a major improvement over other GIS software. The reality with control points is that often you have to use your best guess as perfect info isn't available.


1,774 post(s)
#12-Feb-19 16:41

By the look of the GUI and the disapearance of many selection tools, it sure doesn't look finished. Happy to hear it is advanced. I'm gonna have to read the manual to know how to rename a component and select features...and almost everything else. M9 is powerful of course, but zero intuitive.


5,491 post(s)
#13-Feb-19 08:22

and the disapearance of many selection tools, it sure doesn't look finished.

9 has far better selection than 8. Once you learn to use it, 9 has easier selection than 8, a much wider range of selection capabilities, and within that range much deeper capabilities. The user manual does a good job of covering selection in many settings, and so do some of the videos.

I'm gonna have to read the manual

Yes, of course. 9 is different than 8. How could you learn to use it if you didn't read the manual? Videos help too.

By the way, it is a very good thing that 9 is different than 8. If 9 were not different than 8 you couldn't open 110GB in 1/10th of a second and pan and zoom instantly, like in this video and you couldn't do likewise with bigger vector data either, likein this video.

M9 is powerful of course, but zero intuitive.

9 is exceptionally intuitive, once you learn it. That's a very different thing than being intuitive for a beginner who hasn't learned 9.

The mission for 9 is to provide exceptionally efficient, low-effort, high-productivity, and intuitive workflow for people who have learned to use the product and who use it regularly. The mission is not to create a limited thing which is "intuitive" for beginners who will not invest time into structured learning.

After all, we're all clueless beginners when we start out to learn a big, new, rich and sophisticated package. If we have the common sense to learn the package in a structured way, spending a few days reading the manual, watching videos and trying examples, we only stay beginners for a few days. After that we spend years at the intermediate or advanced level.

There's endless interest in 9 in creating more efficient workflow and more intuitive GUI elements for people who have learned the product. 9 already is far ahead of 8 in that respect, much more intuitive for people who know it.

9 is more orthogonal than 8 ("self similar" as Tim put it very well) with an incredible ability to use and reuse what you learn in one setting in other settings as well. In contrast, 8 is far more special case: you end up having to learn as a special case various ways of doing things depending on their setting, even though the activities are different sides of the same coin. 9 implements such things so they can be learned just once and then done in a much more powerful, faster, and more flexible way in many settings. 8 is better than other "classic" GIS in that respect, but still, it is way behind 9.

By the way, all of the very best applications are like that, Photoshop being a good example. One reason Photoshop, Oracle, Visual Studio, and Mathlab all have a reputation for being wonderful software is that they provide intuitive and effective workflow for people who have learned how to use them, while being criticized by beginners who don't RTFM as not being intuitive.

Once you learn 9, if you think there is anything that could be done more efficiently or anything in the user interface that could make it more intuitive, speak up. Send in a suggestion.

Likewise, once you learn 9, if looking back there was anything you felt would have helped a diligent student learn the package more quickly, such as a missing topic, a video on a particular topic, or anything else, speak up. Send in that suggestion.

I write "diligent student" not because I think everyone should be perfect and never be lazy, but because helping diligent students learn more quickly is a realistic proposition, something that can be accomplished with better topics and videos. In contrast, trying to help people learn who don't read documentation is a much more difficult task. So let's do the reachable goal first.


843 post(s)
#13-Feb-19 14:18

Well it worked for one session with the dll files copied into the manifold folder, but is now back to "Can't establish connection to data source" I don't think I have changed anything, so am now even more baffled than before.

The saved database connections that worked previously no longer connect, and I can't make a new one.

PostgreSQL server still running fine and accessible with manifold via ODBC driver and with PGadmin or QGIS.


5,491 post(s)
#13-Feb-19 16:42

It's best, of course, if you are using 8 to sort out the integration issues using 8. But, just out of curiosity, maybe it would be worth it to try Viewer? Use the portable installation, and put the .dlls for PostgreSQL in the bin and bn64 folders.


843 post(s)
#14-Feb-19 14:47

I skipped over the "are you using 64 bit" in a previous post.

I had copied the dlls from 64 bit postgres installation into my 32bit manifold folder.

Today I have found the 32 bit postgres dlls here

and copied those in instead, and it is working again.

Now I am wondering how the 64 bit dlls worked just once?

Hopefully it will keep working this time!

I have 9 installed at work, but the postgis install at home, so haven't tried with 9, but will as soon as I get around to a postgis install at work.


843 post(s)
#25-Feb-19 08:52

Can now confirm M8 with PostGIS 10 works using the 32bit dlls.

I also had a go with M9 Edge and it seems to connect with no problems without needing to copy the dlls (although they are on my system for M8).

I'm guessing the friendly names etc you can add with Database Administrator for M8 don't work with M9, although I haven't tried reading the help.

Mike Pelletier

1,602 post(s)
#13-Feb-19 22:57

The catch is "... and who use it regularly". It's not too painful to read the documentation but it's a drag when you forget simple things like alt-click to see record info. The shortcut guide is a big help in this regard. It would be nice if the comment component could contain some nice formatting options so one could fill it with a useful guide.

In the future, I hope to use Viewer for non-manual reading people that have only a few basic tasks to accomplish. They open up the project and first go to a relatively eye appealing comment component called "How To". They reread as often as necessary. Perhaps they eventually want a full version so they can do more.


5,491 post(s)
#14-Feb-19 06:40

They open up the project and first go to a relatively eye appealing comment component called "How To".

There's some of that in some of the projects that have been published for Viewer use, which try to hide all the tables in folders, and right at the top level is a Comments component called "Double Click to Read Me" or something equivalent

I think that sort of thing has to be specific to the project. If there is something you publish for non-GIS people to view in Viewer, it probably is intended to be used in a limited set of activities. So, whatever that limited set of activities may be, you can write a "How To" just for those.

But absent a specific, limited set of objectives for a project, what do you do for a Viewer that can do thousands of things, if you don't know which of those thousands are the desires of somebody who has just downloaded it? In that case, the "How to" ends up being the first two chapters in the User Manual, supplemented by a few hundred other topics.


5,491 post(s)
#14-Feb-19 14:56

It would be nice if the comment component could contain some nice formatting options

Sorry I missed that. Might be easier just to have an HTML component that launches in a browser window, or a PDF that launches whatever is the default PDF reader. In principle, there's no reason why a project cannot have components that are containers for various types of media, like MP4 or whatever. Could even launch a video.

[Edit... I sent this in as a suggestion. :-) ]

By the way, I don't disagree it would be handy for Comments to have some more controls and some more formatting options, just for regular ease of writing comments. But if there was a general way to launch a container with stuff in it that Word or something else could create, you'd get a lot of richness without having to re-invent the wheel.

Mike Pelletier

1,602 post(s)
#14-Feb-19 16:39

Launching videos, HTML, or PDFs and tailoring a project to non-manual reading users would be super! Thanks. Some basic rich text formatting for the Comment pane would go a long way to making it better for technical users too.

Perhaps allow some rich text formatting and color background highlighting of the components in the Project Pane. That could dress things up a bit visually and help users find their way around. This would be good for intermediate users that run a full version but still only have a very limited number of tasks. These folks and Viewer users will ask for more eye candy :-)

166 post(s)
#14-Feb-19 17:07

Just simple word wrap on/off would be nice, I'll send suggestion


8,634 post(s)
#19-Feb-19 12:07

Late to the thread...

By the look of the GUI and the disapearance of many selection tools, ...

Which selection tools do you have in mind? I am not saying we have an analog to everything, just wondering whether this is about Select Adjacent / Contained / etc - we added them some time ago, maybe they aren't discoverable easily enough (see the Select pane).

Or do you mean select by polygon? We don't have it yet, true, it's going to be added.


843 post(s)
#11-Feb-19 14:40

You say libpq is in path but it might be better to copy libpq.dll and its dependencies beside manifold.exe (check file properties. Unblock each if necessary).

Thanks, I just copied all the dlls from the postgres installation into the Manifold System folder and it is now working nicely.

Actually got the same answer today from an old internal document I found from when we set up server console, and also the same answer again from Art Lembo on the udemy enterprise gis course forum.

319 post(s)
#08-Feb-19 10:10

try localhost:5432

what version is your libpq.dll? (File Properties -> Details).

Are you running in 64bit?

You say libpq is in path but it might be better to copy libpq.dll and its dependencies beside manifold.exe (check file properties. Unblock each if necessary).

These files are needed:

Latest 9.x and 10.x ()








Latest 11.x









I do not know if 11+ works with M8. 10 should and 9 definitely does.

Links to prepared collections in google drive





3,113 post(s)
#14-Feb-19 03:22

Postgres .dlls

— just putting this comment here for a future search, or if anyone else uses the search bar.

Manifold User Community Use Agreement Copyright (C) 2007-2017 Manifold Software Limited. All rights reserved.