New at 9.3 - Top Ten

We asked the team to give us their Top 10 New Geoprocessing Features in ArcGIS 9.3. The list quickly turned into the top 20, then the top 30, and then we had to stop. It's by no means exhaustive, but we think it includes the major improvements we made that will make your geoprocessing work a whole lot faster and easier.

In no particular order, here's the first top 10 list.

Join Field tool

At last year's UC a common question was how to permanently add fields from one table to another. The solution at the time involved a combination of the Add Join, Add Field, and Calculate Field tools. With version 9.3, the new Join Field tool does all of this in one step, using a fraction of the time. As with Add Join, Join Field accepts two tables, each with a common attribute field. The tool joins the two tables based on this field and you have the option of selecting which additional fields you want to have included in the final output. The Join Field output is a permanent update of the input table or feature class.


See the Join Field  tool reference page for more information.

Error messages

We standardized warning and error messages and gave each one a unique number. So now, when a warning or error comes up on a tool dialog, progress dialog, or command line window, you can click the number to open a detailed description. This description lists probable causes of the error or warning, suggestions to remedy the situation, and links to relevant help topics.

 

See Understanding geoprocessing tool errors and warnings for more information 

Near with lines and polygons

Have you ever been required to find the nearest distances between a mixture of point, line, and polygon features?  This would seem to be a very easy operation. However, until ArcGIS 9.3, it required building a script or program. With ArcGIS 9.3, the Near tool is improved to support all feature types.  In addition, multiple near feature classes can be supplied.

Generate Near Table tool

The Near tool modifies the attributes of the input feature class. This often caused problems when the input dataset was read only. With ArcGIS 9.3, there is a new Generate Near Table tool. Instead of modifying the input feature class as Near does, Generate Near Table creates a new table containing the proximity information. This tool also contains an option to find the distances to all features, not just the nearest.

Performance improvement using ArcSDE data with GpTools

We've made three major improvements when using ArcSDE data with geoprocessing tools:

    By far the most dramatic improvement is the result of better management of ArcSDE connections. Instead of creating multiple connections to ArcSDE for one execution of a tool, we now create only one connection. (This issue is also resolved in version 9.2 service pack 5).

    The performance of the Append tool has improved because the spatial index of the output dataset is created only once instead of many times.

    We made many improvements to our internal topology analysis engine used by overlay tools. The performance of tools in the Analysis toolbox has greatly improved when using large ArcSDE datasets.

Python scripts run in process

In 9.2 all Python [.py] script tools created a separate process and executed outside the application process. The way you noticed this is that it took as much as 5 to 10 seconds for a script tool to start up. In 9.3, we're taking advantage of the application already running. This greatly improves script tool performance in two major areas: Initial start-up of a script tool and the use of cursors within the script tool. It also greatly improves the start-up performance of scripts tools using smaller datasets as inputs.

See Running a script in process for more information. 

Geodesic buffers

In 9.2, the Buffer tool created buffer polygons assuming that all coordinates represented a flat Euclidean space instead of a spherical space, like the Earth.

In 9.3, the Buffer tool will create geodesic buffers if:

    The features to be buffered are points or multipoints.

    The input feature are in geographic coordinates (i.e., longitude/latitude).

    The buffer distance is in Euclidean linear units (i.e, feet, meters).

The output buffer polygons will take into account that longitudinal distance varies varies with latitude, as illustrated below.


 See How Buffer works for more information on geodesic buffers.

Script tool validation

You can now have your script tools behave like system tools (those that come with ArcGIS). You can enable (gray-out) parameters based on values in other parameters, provide keyword lists that dynamically change, put parameters into different categories, calculate default values on-the-fly, and update the description of output data for use in ModelBuilder. These capabilities are all found on the new Validation tab you see on a script tool's property page.

See Customizing script tool behavior for more information 

Spatial Statistics: Ordinary Least Square Regression and

   Geographically Weighted Regression Tools

The Spatial Statistics toolbox has powerful pattern analysis tools to help answer "Where?" questions like:

    Where are people persistently dying young in the United States?

    Where are our kids consistently turning in high test scores?

    Where do we seen an unexpectedly high number of traffic accidents?

With ArcGIS 9.3, new regression analysis tools are added that allow you to answer the next logical set of questions relating to "Why?" like:

    Why are people persistently dying young in particular places across the United States?

    What are the factors contributing to consistently high test scores?

    Why are there so many traffic accidents in particular hot spot locations?

Ordinary Least Squares Regression (OLS) is a commonly used global linear regression method used to generate predictions, or to model a dependent variable or process, in terms of its relationships to a set of explanatory variables. It creates a single equation to represent those relationships. OLS is the starting point for all regression analysis (including all spatial regression analysis). It often breaks down when used with spatial data, however, in the case where the relationships being modeled change across the study area. An education variable, for example, might be a good predictor of crime rates in one part of the study area, but may not be significant in another part of the study area. Geographically Weighted Regression (GWR), on the other hand, allows data relationships to vary. GWR is one of several spatial regression techniques increasingly used in geography and other disciplines. It provides a local model of the variable or process you are trying to understand/predict by fitting a regression equation to every feature in the dataset. GWR constructs these separate equations by incorporating the dependent and explanatory variables of features failing with a particular user-specified bandwidth of each target feature. ArcGIS 9.3 includes "Regression Analysis Basics" and "Result Interpretation" documentation with strategies and graphics aimed at helping the user apply these methods effectively.

See An overview of the Modeling Spatial Relationships toolset for more information. 

Batch Grid: deferred validation

If you use a tool in batch mode, you may have noticed that filling in the batch grid became slow if more than 10 rows or so were entered. This was due to the fact that every value in every row was validated whenever you clicked the mouse in a cell.

In 9.3, you control when validation occurs by click the Check Values button in the lower right of the batch grid, as illustrated below.

 

Validation always occurs when you click the OK button. 

See Using the batch grid control for more information. You can find the discussion of the Check Values button at the bottom of the topic.  

Published Tuesday, July 29, 2008 9:44 AM by dmhoneycutt

Comments

No Comments
Anonymous comments are disabled