Clicking on Edit View in the database structure tab opens a new SQL tab
with a DROP VIEW statement to delete the existing view and a copy of the
CREATE VIEW statement used for the current view. When the user modifies
the CREATE VIEW statement introduces an error in it, and then executes
the statements, the view is deleted and an error reported. Fixing that
error and re-running the script then fails at the DROP VIEW statement.
This means it makes more sense to make it a DROP VIEW IF EXISTS
statement.
See issue #2661.
It does not seem like there is a good reason to always refresh the table
browser dock (there is always exactly one at this point) when opening a
database file, even when it is not visible. This only generates
unnecessary queries to the database.
Remove the F5 shortcut for refreshing the DBHub.io remote databases
because it is ambiguous and can make the main view not work as expected.
Since the refresh action is not super important in the DBHub.io dock any
way it is probably best to just remove the shortcut there.
See issue #2695.
When multiple cells are selected pressing the Delete key tries to set
all of them to an empty string. In case of a unique constraint or
similar constraints this throws an error. This commit copies the
behaviour of the set to NULL menu action in that it aborts at the first
error to avoid multiple error messages.
See issue #2704.
This fixes a possible crash when trying to remove a table constraint
from a table definition. The issue was introduced in commit
4e1d1ff49f.
See issue #2670.
Support of extended selections (non-contiguous cell selections using
Control+Click). The following changes were necessary:
- Copy to clipboard iterates over rows and columns and leaves holes for
non-rectangular selections as empty cells (HTML) or as NULL (SQL).
Fix: Additionally the SQL copy uses NULL when the cell has NULL, not ''.
- Shortcuts for column and row selection take into account non-contiguous
cells and only select those columns/rows.
- The legend in the status line counts correctly non-contiguous rows or
columns.
- Delete Record has been adjusted so only contiguous selected cells are
removed in a single step. Fix: selecting line after removing has been
deleted since the standard behaviour is giving better result.
- ExtendedSelection has been enabled in TableBrowser, it was already
enabled in Execute SQL table-widget.
Possible improvements: pasting from the internal clipboard does not keep
the layout of copied non-rectangular selections.
See issues #1104 and #2638
Since in any case, the empty file has nothing to import and it's admitted by
SQLite as a valid database, it's better to revert to the previous behaviour,
where an empty file dropped or opened by the file manager was treated as a DB
and indeed worked.
The behaviour was changed by cfa4dee3be
See issue #2636
The Clear Filters button is enabled if and only if there is some filter
set for the browsed table.
The Clear Sorting button is enabled if and only if the table is sorted by
some column.
See issue #2616
"Yes" to not be prompted next time for the same table, but yes for different
tables.
"Yes to All" to not be prompted again for any existent tables.
See issue #2633
* Add auto completion for math functions
Math functions are introduced from sqlite 3.35.
Refer https://www.sqlite.org/lang_mathfunc.html
FYI, it need to compile sqlite with -DSQLITE_ENABLE_MATH_FUNCTIONS
* Add auto completion for math function when it had been enabled
Math function can be used when sqlite is equal or higher than 3.35 and
compiled with SQLITE_ENABLE_MATH_FUNCTIONS. When sqlite is compiled with
SQLITE_OMIT_COMPILEOPTION_DIAGS, there is no way to check the compile
options. In this case, we'll check only the sqlite version. Otherwise,
it will check whether sqlite was built with SQLITE_ENABLE_MATH_FUNCTIONS.
This commit replaces the regular expressions for removing end of line
and block comments from SQL queries provided by the user by a hand
written state machine. This makes the code a lot faster, especially for
longer SQL scripts with many statements and many comments in them. It is
also not harder to read than the rather complex regular expressions from
before - possibly even easier to read.
See issue #2619.
In this case too, the corresponding menu entry for exporting the query
results to JSON was missing although it was straight forward to add it
with the same scheme as for CSV.
Renamed the CSV action for consistency.
Related to issue #2607
This check was supposed to help the user understand what this operation was for,
but it was actually preventing to save as view table states that had global
filter, display formats or custom order and not column filters. It's better to
give the user the freedom, even to create maybe useless vies as:
`CREATE VIEW "example_view" AS SELECT * FROM "main"."example"`
See issue #2615
This fixes a bug introduced in 73efa11680.
Because SQLite reports ALTER TABLE statements to return one column worth
of data, DB4S assumed they are close to a SELECT statement and therefore
did not fully execute them.
See issues #2563 and #2622.
The corresponding menu entry for exporting the query results to JSON was missing
although it was straight forward to add it with the same scheme as for CSV.
Related to issue #2607
Support these settings from the command line:
--option importcsv/separator=\, --option importcsv/quotecharacter=\'
Since they are stored as string, they cannot convert to QChar unless
manually converted.
See issue #2589
This will allow associating DB4S to CSV files and dropping a CSV file to be
importing into a new in-memory database, when there is no DB open yet.
See issue #2589
The argument can be passed several times and all the CSV files are added to
the Import CSV dialog. When no database to open is passed in the command line,
the CSV files are imported into a new in-memory database, which could later
be saved as a file, if desired.
This option could be used as basis for adding a file association to CSV files
for DB4S.
See issue #2589.
- Command line arguments are not given to translations, only descriptions
and placeholders.
- Formatting is done programmatically, so developers and translators don't
have to adjust that themselves.
This will affect translations but will be better in the long run.
This fixes some severe bugs in the Browse Data tab with editing and
deleting rows in WITHOUT ROWID tables.
These were introduced in 02db68107a.
See issue #2582.
This improves the performance of executing multiple modifying SQL
statements, like INSERTs or UPDATEs, in the Execute SQL tab a lot by
not updating the plot dock after every single statement.
See issue #2572.
This commit does a lot of refactoring. But most noticeably it changes
two things:
1) Instead of saving all objects (tables, views, indices, triggers) of a
schema in a common map, we now store tables/views, indices and
triggers in three separate maps. This has a number of benefits:
- It resembles more closely how SQLite stores its data internally and
therefore achieves greater compatability e.g. for databases with a
view and a trigger with the same name.
- It reduces the need for runtime polymorphism. This makes the code
run a bit faster.
- By working with explicit types more often more error checking can
be done at compile time, making the code less error prone.
- The code becomes a lot clearer to read.
2) By making View inherit form Table, views are now a sort of tables.
This has the following benefits:
- This is a again how SQLite stores views internally which again
should increase compatibility a bit.
- We mostly treat views and tables the same anyway and with these
changes we can unify the code for them even more.
For the same reasons we changed the ORDER BY part of query objects to
use their names instead of their indexes to address columns, this commit
changes the handling of the WHERE part of a query to use column names
too.
This changes the structures for representing SELECT statements to save
the names of sorted column instead of their indexes. This change has
several benefits:
- It prepares the Query class to store actual real-world SELECT
statements in the future.
- It prepares us to sort by expressions instead of just columns.
- This way we do not need an extra list of column names to generate the
ORDER BY part of the SELECT statement when building it.
This simplifies the usage and the implementation of the SqliteTableModel
a bit by making the two operating modes it supports (manual query and
interactive query) more obvious in its public interface as well as
simplifying the control flow in the private implementation.
Change some function parameters which are named like "dont..." to make a
positive statement. This should hopefully avoid some confisions with
double negation.
The feature is implemented using the SQLite backup API.
https://sqlite.org/backup.html
This allows saving in-memory databases and saving database files to another
file name.
When opening a database and clicking on a table in the Database
Structure tab, the buttons for modifying or deleting the table were
still disabled. Only on the second click activated the buttons
correctly.
See issue #2528.
Martin Kleusberg