The file test.csv (attached if possible) is a csv presentation of a single-column spreadsheet with 1048575 rows such that the first row contains only the letter C and every other row contains only an integer number in the range from 6 to 3145725. For me, OpenOffice 4.1.1 imports test.csv successfully in less than one minute whereas IBM Rational DOORS Client 9.5.1.1 is practically unable to import test.csv. I have seen IBM Rational DOORS Client 9.5.1.1 importing a considerable amount of rows in a couple of weeks, but the network connections that I have used are such that IBM Rational DOORS Client 9.5.1.1 typically in less than a couple of weeks loses a connection either to the DOORS server or to the license server, even in such a way that there is no way to save any of the already imported rows. In my import experiments, IBM Rational DOORS Client 9.5.1.1 has consumed much more memory than I would expect on the basis of the 8018232 bytes size of test.csv). I am just wondering whether there is a way to configure IBM Rational DOORS Client 9.5.1.1 to be able to import test.csv successfully in less than one week. I do not know any industrial project that would be interested in having a DOORS module with more than a few thousand objects. However, an industrial database can well have millions of items. kvarpaan - Thu Sep 11 02:49:53 EDT 2014 |
Re: Spreadsheets with very many rows are very slow to import Hi You should become familiar with memory leaks in doors dxl and with the handling of the eval_()-function. Then you shall realizes, that a csv-file is a text file, wich can be read in line by line. Read in line by line, avoid memory leaks or get them rid by using eval_() and hope, that doors will be able to hanle a module providing one million objetcts. The alternative: Doors is a requirement tool. Do you really think that you are dealing one million requirements? Do you really think, that doors is the correct tool? Best regards Wolfgang |