On 2014/12/13 21:46, James K. Lowden wrote:
So the number of tools with feature X is no measure of the value of X. (Notable example: the tool should keep every query and result in a time-sequenced transcript log, so that prior results can be re-examined and prior queries modified. Most tools disassociate query from output and invite the user to modify the query in-place, destroying the prior.)
This is hardly a function of the DB admin tools though, it's a research function or application function (depending on whether the inquest is theoretical or physical). That said, many of the tools I know do allow saving queries as scripts with the possibility to examine output at every step, but not all of them, so this might be somewhat valid. More importantly, I was not touting the idea that because all the tools use feature X, it is therefore valid... I more conceded the fact that the request for feature X is valuable (mostly) only to those tools, which probably renders it less urgent - quite in agreement with your point.
My first question, then, is whether or not the rowcount is so interesting that it must be known before a table can be operated on. I suggest the answer is No. The relative & approximate sizes of the tables is known to the admin in most cases and, when it is not, the information is readily discovered on a case-by-case basis. Would a proxy figure do? Is it enough to know the number of pages or bytes allocated to a table? I don't know if such is available, but if it is perhaps that would serve your purpose.
Yes, this would actually do, but it is not available as you rightly pondered. To the point of necessity, I have to disagree. It is nearly always the first thing I want to know. When someone here is troubled by a query running time... first question is: how many rows are in which joined tables? To state the problem a bit simplistic - It is hard to fathom the meaning of O log N without a clear understanding of what both O and N might be. And that's just from a DB admin perspective, in companies where the business analysis data matter, lots of queries are usually stored as tables for further analysis, and the first thing asked is: How many? Other times that figure serves probably only as bemusement to big data fans. It's usually (I'd say vast majority of cases) an easy and fast step to ascertain though (as this discussion pointed out) hence me resting the case - but I do stand by the point that the need isn't invalid.
_______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
