SQL console: room for improvement
The new SQL console in IDEA is a nice addition. However, the current state is quite raw and unpolished.
While having familiar completion for columns etc is very nice, I'm still using external tools at the moment for any SQL querying.
Let me try to compile a list of "blocking" problems:
1) Console in implemented as a pane tab
Whatever type of text I edit, I expect to do it in the editor pane of the IDEA.
Having it in a small tab at the bottom sucks, since:
a) I'm limited to this really small region of my screen
b) When looking at query results, the effective "editing area" is reduced even further
I'd expect to edit SQL in a regular editor tab. This is where I perform all of my other editing after all.
The results of a query execution, however, would fit perfectly well in a bottom-docked results pane.
Of course there's the question of "but is it really an editor file?".
I don't care - make it some kind of virtual file, perhaps showing the name of the datasource in the editor tab?
2) Results need a table
Please present results in a simple table component. 'Nuff said.
3) Completion for tables/columns
Having completion for all the functions supported by my current SQL dialect is nice.
However, most of them (>90%) I will invoke very rarely.
At the moment the completion suggestions for functions easily dwarf/pollute the completion suggestions for table/column names.
Here's an idea: provide SQL function completion only after invoking code completion the second time.
Similar to the new extended smart completion in Diana, you could initially show a message like "Invoke code completion again to receive SQL function suggestions".
Of course I have tons of other requests and ideas, but I think above 3 are simply critical for an acceptable user experience.
请先登录再写评论。
I just tried "Refresh Tables" again using 8.0.0 build 9013, and the problem remains. After 1 minute, it says it has loaded 4% of my tables, so I guess it will take 25 min or more to show me a list of tables?
As I mentioned before, DBVisualizer loads the list of tables in 2 secoonds. That is several orders of magnitude faster.
Attachment(s):
9013_database_loading_table_structure.png
Interesting.
I'd guess there are two approaches to retrieving the table metadata:
1) Using the JDBC DataBaseMetaData
2) Using information schema (if available)
It should be pretty easy to figure out which approach is used by IDEA vs DBVisualizer.
Another comparison is SQuirrel SQL.
Since it is open source too you can easily work out which it uses!
Thanks for mention of Squirrel SQL. I downloaded and connected it to the same Sybase ASA db server and it took 5 seconds to display the navigation tree compared to 2 seconds for DBVisualizer.
This Squirrel SQL is looks very full featured also; I might even try using it instead of DBVisualizer. From looking at the screenshots it might have more features than DBVisualizerFree version. The developers Colin Bell, Gerd Wagner, Rob Manning, et al. did a nice job. It even looks like they used IntelliJ to write it!
Hi Gregory,
Will this be fixed in 8.0.x / 8.x?
I *really* like the SQL console for the very exact completion if offers, but the superfluous indentation is a huge annoyance.
Taras
It will be.
Taras Tielkes wrote:
--
Best regards,
Maxim Mossienko
IntelliJ Labs / JetBrains Inc.
http://www.intellij.com
"Develop with pleasure!"
Any - tentative - date for the fix?
As promising as the JDBC console is, it's completely useless because of this problem.
-tt