ora | An Oracle database driver for the Go programming language | Database library

Sorry, there's no option for this.

I'll investigate an ER to increase this to 4000.

As a workaround you could have scripts for testing your procedures, and then run those exactly as you please in the SQL worksheet using the Execute as Script (F5) button.

There are two issues here:

1. Oracle Database is not supported on ARM processors, only Intel. See here: https://github.com/oracle/docker-images/issues/1814
2. Oracle Database Docker images are only supported with Oracle Linux 7 or Red Hat Enterprise Linux 7 as the host OS. See here: https://github.com/oracle/docker-images/tree/main/OracleDatabase/SingleInstance

Oracle Database ... is supported for Oracle Linux 7 and Red Hat Enterprise Linux (RHEL) 7. For more details please see My Oracle Support note: Oracle Support for Database Running on Docker (Doc ID 2216342.1)

The referenced My Oracle Support Doc ID goes on to say that the database binaries in their Docker image are built specifically for Oracle Linux hosts, and will also work on Red Hat. That's it.

Linux being what it is (flexible), lots of people have gotten the images to run on other flavors like Ubuntu with a bit of creativity, but only on x86 processors and even then the results are not guaranteed by Oracle: you won't be able to get support or practical advice when (and it's always when, not if in IT) things don't work as expected. You might not even be able to tell when things aren't working as they should. This is a case where creativity is not particularly rewarded; if you want it to work and get meaningful help, my advice is to use the supported hardware architecture and operating system version. Anything else is a complete gamble.

Not like that, but

It's easy to create a Pandas dataframe from a Python dictionary that contains lists. For example, the following code creates a dictionary containing ticker symbols, bid prices, and ask prices:

You are calling a PL/SQL function that returns a PL/SQL collection type (both defined in your package) from a SQL context. You can't do that directly. You can call the function from a PL/SQL context, assigning the result to a variable of the same type, but that isn't how you're trying to use it. db<>fiddle showing your set-up, your error, and it working in a PL/SQL block.

You could declare the type at schema level instead, as @Littlefoot showed:

There are several aspects here ...

First: the optimizer is free to apply the predicates in any order. In your case, it looks like it applies the spatial filter first, then applies the selector on number of vertices. Which means you get the exception before the test on number of vertices.

That the optimizer does this is natural: changing the order would mean a full table scan to only return the geometries with 4 points or more, then pass the result through the spatial filter. That would be very slow, and the optimizer rightfully prefers using the index first.

There is no mechanism (hints or otherwise) to control this behavior. Using a subquery or a view will not make any difference: the optimizer will flatten the query into a simple one. Possibly a subquery with a NO_MERGE hint could work: but it would have the above effect of forcing a full table scan and a full pass of all geometries through the spatial filter. Not a good thing.

Second: polygons with less than 4 vertices are incorrect. The simplest polygon is a triangle. It has three points (A-B-C), but all polygons must close, i.e. be encoded as four vertices: A-B-C-A. That is one of the rules defined by the OGC Simple Features for SQL specification. There are others that polygons must adhere to:

• Absence of redundant vertices
• Orientation (counter-clockwise for outer rings, clockwise for inner rings, i.e. holes)
• Absence of self touching rings
• Ordering of the rings (an outer ring must be followed by its inner rings)

Shapes that do not adhere to the rules are invalid. What happens when you use invalid shapes is actually undefined. Depending on the nature of the error and the action you do on this shape (query, measure, buffer, clip, merge ...), you may get any of the following behaviors:

1. the error is ignored and you get a correct result
2. you get an exception (that is your case)
3. you get no error, but the result is incorrect

The worst possible outcome is #3: you cannot trust the results of your application. It may return the wrong area in m2 of a parcel. Or it may say that two adjacent parcels do not overlap, when in reality they do ... This is very bad.

Data quality is of prime importance when manipulating and processing spatial data. Note that errors are generally not visible: most map-mapping tools are resilient enough to still show the shapes, and defects are for the most impossible to detect visually.

The solution is simple: make sure your data is valid. For that you can use the SDO_GEOM.VALIDATE_GEOMETRY_WITH_CONTEXT(). Run it over each shape. It will tell you which shapes are incorrect, and what the error is.

There is also SDO_UTIL.RECTIFY_GEOMETRY(). This one will attempt to correct the most common errors:

• Removes redundant vertices
• Reorients and re-orders rings
• Corrects some self-orientations

It does not correct the errors you see (less than four points) because it is uncertain of what is actually wrong. You need to look at what those shapes are, and more important where they came from. Then either correct them or remove them.

Based on that WSDL file, a code like this:

We encountered the same issue on one of our Oracle 19c databases after applying the Oct2021 Release Update. We were able to solve the problem by backing out the RU. Since we wanted current security patches, we applied the Jul2021 RUR1 in its place and re-tested. The problem did not recur after after the RUR was applied, suggesting that the issue was caused by one of the non-security bug fixes or minor feature updates included in the RU.

For those unfamiliar with OraCorp's RU/RUR patching strategy, it's important to understand that July 2021 RU Revision 1 contains all of the latest security fixes as of Oct 2021. For details, please review MOS Doc #2285040.1, which explains how Release Updates and their Revisions are related. I'll note that while our team has maintained documentation on deployment of each quarter's RURs, this is the first time we've actually had use for a Revision.

I'd never (based on this answer) recommend this one step approach where the dbWriteTabledirectly maintains the target table.

Instead I'd recommend a two step approach, where the R part fills a temporary table (with overwrite = T i.e. DROP and CREATE)