The most obvious difference between direct ODB access (Attach method) and results translation (Translate method) is that the results are not imported into the Patran database in the case of the former, while they are for the latter. Direct access avoids redundancy and saves disk space, while Translation uses more disk space, but takes less time to retrieve results for postprocessing.

The following sections describe other differences that users should be aware of, before deciding which method to use.

The names used for the result types within an ODB attachment come directly from the field output description fields of the ODB database. Using the “direct access” philosophy of bringing the data in as-is, there is no attempt to map those names to the same names used by the Translate method (listed in Table 4‑1).

Therefore, direct ODB access will use Abaqus terminology exclusively in generating the result type names. The primary name is equal to the field output description field, while the secondary name is the field output key. For example, the stress tensor result type is “Stress components, S”, where “Stress components” is the field output description, and “S” is the field output key.

For results such as reaction moments or rotational displacements, the ODB database saves space by only storing results for the non-zero component, whenever possible. So, if non-zero values for moments only occur in the Z component, then the ODB database stores it as a scalar result (e.g. key RM3). However, the Translate method will import the results as vector results, with the X and Y values always being zero.

This difference may cause confusion when comparing translated results against direct ODB access via the quick or fringe plot operations, where reaction moments and rotational displacements are concerned. The default “invariant” for fringe plots of vector data is “Magnitude”, which is always a positive value. If the magnitude of the translated vector data is compared against the ODB scalar data, then they will not always match (all negative data from the ODB access will be flipped positive in the translated plot). To compare “apples with apples”, one must display the appropriate component (Z from our example) from the translated case, and compare that against the scalar (key RM3) from the direct ODB access case.

During translation, only non-zero reaction force data is imported. Direct ODB access, on the other hand, returns zero vectors for any nodes that do not have any reaction forces. This makes no difference for the display of reaction force vectors; however, if one displays a fringe plot distribution of the reaction forces, the fringe plots vary between translation and direct ODB access dramatically. The translation plot is all black, with only the min/max values displayed on a hidden line plot; while the ODB fringe plot shows a color distribution from the zero values (white over most of the model) to the non-zero values. For the latter, the contours only vary over elements with nodes having non-zero reaction forces.