Problem Note 69158: Databricks tables created with SAS/ACCESS® Interface to Spark might create columns with different lengths after a period of inactivity
 |  |  |  |  |
If a SAS/ACCESS Interface to SPARK libref is assigned to DATABRICKS and used immediately, it might describe columns of a different length than if the libref is assigned but not used for five or more minutes. For example, if a table is read immediately after assigning a SPARK library, a CHAR(8) column has a length of 8 in SAS. If the session is left idle for five or more minutes, a rerun of the same code might yield a column of length 255 in SAS for the same CHAR(8) source column.
This issue occurs because the DESCRIBE FORMATTED statement to retrieve table metadata fails due to a time-out within DATABRICKS after the five-minute wait period, and the fallback behavior yields an incorrect length.
Click the Hot Fix tab in this note to access the hot fix for this issue.
Operating System and Release Information
| Product Family | Product | System | Product Release | SAS Release | ||
| Reported | Fixed* | Reported | Fixed* | |||
| SAS System | SAS/ACCESS Interface to Spark | Linux for x64 | 9.4 | 9.4 TS1M7 | ||