In your prior message you indicated that the features selected are done so without benefit of common attribute (or set of attributes) or spatial condition that can be expressed as an algorithm (SQL phrase). That is, you selected them "manually" using your own heuristics not represented by anything in the database.
In affect you, in your own mind, have created some "new" common binding logic between these features. You are now left with two methods for retaining these groupings: 1) The quick and dirty method 2) Modifying your data structure (and data) to implement the selections. 1) The quick and dirty method (which you said you do not want to do, and I can agree with for a number of reasons, but it does work) is to "Save" the grouping as a collection in a file. If you do create new tables consider using the metadata capabilities to identify / manage the information, or create your own file scheme to do the same. a) Save the selection as a table, and deal with it. b) Save the selected identifiers, and possibly the geometry as a new table (a table with only identifiers will do). This only reduces the amount of information redundantly stored. You can always "bind" the selection list back to the original list to gain access to the attributes. What you have done is implicitly created a new attribute. 2) Modifying your data structure (and data) to implement the selections. If the process is supposed to be repeatable by yourself or others then I would suggest that you have a data problem. It sounds like there is a missing attribute and/or spatial features (layer) that would provide the criteria to select on, but you don't have. The solution is to create a new attribute (column in the tale) that represents the value or logic that you want to select on. Then the selection can be saved / used as a SQL phrase (it may be more than one attribute). In terms of a process you can implement a simple or more thorough solution to your problem. Simple method: (This may just be a temporary step until you can do better) Modify your table structure, adding a new attribute that represents your selection logic. Make your selection as you have in the past, and then insert (update) the new attribute you created with some meaningful value. From then on you can use an SQL phrase. The thorough solution: I would encourage you to go through a design process to determine the data requirements: Business logic -> data flow -> data definition and normalization exercise to determine what characteristics are needed for your application. You must be able to express you logic for selections as an algorithm. Don't forget that the algorithm you use for your selection can contain both attribute conditions and spatial conditions. It may be that you don't have attributes in an existing data set, or a spatial dataset of boundaries (or whatever) you need is not present. In either case you need to determine your data requirements and how you can obtain or create them. The solutions may be simple ones as suggested above, or more complicated (involving new data collection). The benefit of the design / specification process is to identify the needs and document any decisions / trade-off that you make. Guy Groves GRG Consulting 6162 S. Annapurna Dr. Evergreen, CO 80439 [EMAIL PROTECTED] -----Original Message----- From: Phillips, Frank [mailto:[EMAIL PROTECTED] Sent: Monday, July 14, 2003 10:31 AM To: 'MapInfo-L' Subject: MI-L Convert Selection to SQL Is there some way to "convert" a selection to SQL so that it can be "saved" in the Workspace? Frank Phillips Manager of Marketing GIS Vulcan Materials Company (NYSE:VMC) Birmingham, AL, USA --------------------------------------------------------------------- List hosting provided by Directions Magazine | www.directionsmag.com | To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] Message number: 7627 --------------------------------------------------------------------- List hosting provided by Directions Magazine | www.directionsmag.com | To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] Message number: 7645
