Problem Note 71031: Performance improvements to SAS/ACCESS® Interface to Google BigQuery
Several changes have been made to SAS/ACCESS Interface to Google BigQuery to improve performance.
- A change was made to how SAS® interacts with Google's Storage API when you are using the READ_MODE=STORAGE LIBNAME or data set option. This change has shown a significant increase in read performance.
- The additional COMPRESS, TRANWRD, INDEX, BYTE, and COALESCEC functions are now considered for push down to the database.
- You can use the DBSASTYPE= data set option to override the MAX_CHAR_LEN= or SCANSTRINGCOLUMNS= options. This change enables you to use SCANSTRINGCOLUMNS=no, set a generic MAX_CHAR_LEN= value, and override the lengths with DBSASTYPE=.
- The NUMERIC data type is now converted to a DOUBLE directly rather than being loaded first as a TKTS_NUMERIC, which is then converted to a DOUBLE in SAS. A new LIBNAME and data set option, FETCH_NUMERIC_TYPE=, enables you to revert to the previous behavior. If you intend to preserve more than 15 digits of precision by converting the NUMERIC to a character value, select NUMERIC for this option. The default is FLOAT64.
Click the Hot Fix tab in this note to access the hot fix for this issue.
Note: The upgrades were made to the SAS® Viya® Platform Stable 2024.09.
Operating System and Release Information
SAS System | SAS/ACCESS Interface to Google BigQuery | Linux for x64 | | |
SAS System | SAS/ACCESS Interface to Google BigQuery (on SAS Viya) | Linux for x64 | Viya platform | |
*
For software releases that are not yet generally available, the Fixed
Release is the software release in which the problem is planned to be
fixed.
Type: | Problem Note |
Priority: | high |
Date Modified: | 2024-11-12 16:20:23 |
Date Created: | 2024-10-31 15:09:18 |