Salesforce record ids are case-sensitive values. However, MS-Access string comparisons are typically not case-sensitive.
Because of this, record ids exported from SF cannot be reliably used for table joins within MS-Access databases. Although it is rare to encounter problems, there is a clear possibility of incorrect results.
We typically use a VisualBasic function to create hexidecimal conversions of Salesforce Id values, guaranteeing that equi-joins always return the correct result.
As of early 2018, MS-Access limits the size of individual databases to 2Gb.
We have observed problems sometimes occur when MS-Access databases reach
sizes below this value - typically around 1.6Gb. Because MS-Access
databases do not relinquish
space unless the user performs a compact and repair
operation, even
the act of deleting and reloading the same data repeatedly during development
can cause problems.
As a result, we typically use a linked system of MS-Access databases during data transformation work:
testand
finaldata loadings within a single project