Hi, yeah, been thinking abou that one... creating the tables is a good
idea, the trick is to keep them identifiable for the consumers of this
info... from one churn of the automation to the next, all tables are
the same in name and appearance. So I was thinking, perhaps we can
trick the user (wo will want to search by build_version number, by
forcing their query to resolve that version number to the tables' long
md5-ish name. So, thinking about a glossary (really a dictionary). I
haven't yet tried it but, something like this should return the
tableName to a user using the build_number instead of the table
name...
glossary_changes = {<build_version> :
c8b669d15150d7109e5f7ab36744a5b7_changes.table}
...assume tableName is 'changes'
resolvedTableName = getTable(myBuildPath, table2Query,
myBuildVersion)
def table_exists(folderPath,tableName):
try:
tableName = os.path.join(folderPath,
'{0}
_{1}.table'.format(md5_hash(self.db._uri),tableName))
return False
except:
return False
def getTable(folderPath, table2Query, ver):
gloss = 'glossary_{0}'.format(table2Query)
tables = gloss.keys()
for mTbl in db.tables:
for tbl in tables:
if mTbl.startswith(tbl):
for row in
db(db[mTbl].buildVersion==ver).select():
if table_exists(folderPath, row.tableName) is
True:
return row.tableName
should return ==> c8b669d15150d7109e5f7ab36744a5b7_changes.table
Thanks for your input!
Mart :)
On Feb 2, 5:21 pm, howesc <[email protected]> wrote:
> i don't know sqlite. never really used it so i can't say much about that.
>
> can you take your file paths and use that to prefix the table names? there
> is nothing stopping you from creating the table name dynamically and
> accessing using db['table name string'].
