Converting an existing DOS relational database to OutSystems

I have a DOS-based relational database and I want to convert this to an OutSystems application. The database engine is C-tree. The program is low-code, based on the block diagram of the required program. There are windows (UI), files, items (temporary storage), and reports (screen or printouts). These are linked with steps (10) to create the program. eg a Dialog step will display a window to enable user input, a Display step will provide information to the user (eg printing in progress), a File step will allow a file to be created, deleted, imported, or exported, etc.

If there is some way of accessing this C-tree data directly that could be useful.

If not, then the data in each table can be exported (in a number of formats, eg comma delimited). I see there is information on importing to OutSystems from an Excel file, and a comma-delimited file exported from C-tree can be imported correctly into Excel.

The C-tree files have linking fields to create relationships between files. The block diagram below shows the basic file structure and their relationships. There are no automatic default index fields, they have to be specified.

The first question is whether the existing data can be pulled from the C-tree database and imported into OutSystems, keeping the current relationships.

The current webinars show how to import one table from an Excel file, however, I don't see any help on how a relational structure can be created from a number of Excel files. So that would be question two if the response to question one above is positive.

Rank: #17

Hi Terry,

First a silly check; is the amount of data in this database so large you would not be able to manually retype it into the new OutSystems application in less then let's say a few days?

Why I ask, that is probably the time it takes to code the conversion.

If you could flatten the database structure in to one very redundant row of comma separated file then you would still have all the data related to each other.

One row of the CSV, can then easily be processed in all the seperate entity records crating the correct relations between them.



Rank: #59350

Hello Daniel,

Thank you for your response. I have done some analysis to see what the implications of your suggested approach would be, and have found the following:

The database files are in two parts, the data records and another file with the defined indexes (The C-tree engine requires at least one defined index). The volume of the data files is 16Mb+, which would be too big for the DOS environment. The second restriction in creating one file of this data is that the C-tree engine has limitations on the number of fields that can be included in one .exe file, and the number of fields required for a single file would be several times more than this limit. A third restriction would be the size of the generated file. I have already had to split one of my files into a "current records" part and an "archive records" part, and each of these data files is nearly 2Mb in size. (I am talking about 20 years worth of data in these files). The "current records" file contains 7660 records, so re-typing would be a mammoth task.

So I am back to exporting table by table, importing to an Excel file, then importing into an OutSystems structure.