|
Changes The table window allows changing the alignment of field values using the Align Left / Align Center / Align Right commands in the context menu for a field. The default size for binary fields in a table window is forced to be constant (was adjusting to the system font). The table window allows changing the format of field values using the Style command in the field context menu. The command is disabled for text fields (this will quite likely change in the future). Fields of all other types support it. The command invokes a different dialog depending on the field type: - binary: select style from the list of available styles, specify language
- boolean: select style
- date: select a standard style or compose a custom style from parts, specify language
- geom: select style, specify language
- number: select style, specify the number of digits / decimals, specify language
- tile: select style, specify language
- uuid: select style
- vector: select style, specify the number of digits / decimals, specify language
All styles and style parts show what kind of output they are producing in the dialog. The language is specified using an editable dropdown box which allows either typing the desired value directly or picking it from several default choices (neutral language + languages for available localizations). There are two cases where the table window ignores field styles: Editing a cell always uses the default style. Otherwise it is too easy to lose data. For example, suppose that a numeric field is set to show two decimal digits. This shows 3.14159 as '3.14'. If we start editing the value and the edit box opens with '3.14', pressing Enter will accept '3.14' and change the value from 3.14159 to 3.14. This is unwanted. To prevent that, when we start editing the value, the edit box opens with '3.14159', showing all available digits. Filtering records by right-clicking a cell and using the commands in the context menu always reports the value of the clicked cell using the default style. Similarly to editing, applying a style would miscommunicate what the filter is going to do. (Fix) The table window no longer hides the euro sign and several other Unicode characters not recognized as printable in the default CRT locale. The Add (Field / Index / Constraint) and Add Identity commands in the Schema dialog focus the last added schema item. The Schema dialog includes a filter box to filter schema items using their name. Filtering by name automatically disables the Add (Field / Index / Constraint) and Add Identity commands to prevent newly added items from being immediately hidden due to not passing the filter. Editing an existing item and changing its name lets the item stay visible even if the new name no longer matches the filter. The Schema dialog includes a filter button with a dropdown menu to filter schema items using their type (All / Fields / Indexes / Constraints). Filtering by type automatically disables the Add (Field / Index / Constraint) and Add Identity commands for items that will not pass the filter. (So, for example, if you set the filter to only show fields, you can still add new fields, but you cannot add indexes and constraints until you change the filter.) The Move to Top / Move Up / Move Down / Move to Bottom commands in the Schema dialog are removed. (These commands could only reorder new schema items and such reordering only made sense for computed fields referencing each other. We felt this was a case where the commands were trying to help those who already knew what they were doing at the cost of confusing everybody else, so we removed them. In the future we will also be able to handle computed fields coming in an arbitrary order, provided there are no circular references.) New query function: StringDateTime. The function takes a string and attempts to parse it into a date time value according to the specified language. New query function: Random. The function produces a random floating-point value between 0 and 1, with 1 excluded. New query function: RandomInt. The function takes an integer limit and produces a random integer value between 0 and limit-1. New query functions: RandomRef / RandomIntRef. The functions produce random values similarly to Random / RandomInt, but take an extra argument of an arbitrary type to use as a dependency. New query function: RandomHash. The function takes an arbitrary value, composes its 64-bit hash using FNV1a, and turns that hash into a pseudo-random floating-point value between 0 and 1, with 1 excluded. (The purpose of this function is to produce a value that is both uniformly random across the domain of argument values, and coincides for calls where the argument value is exactly the same.) New query function: RandomHashInt. The function takes an arbitrary value for the hash and an integer limit, composes the 64-bit hash using FNV1a, and turns that hash into a pseudo-random integer value between 0 and limit-1. New query function: DataHash. The function takes an arbitrary value and composes its 32-bit hash using FNV1a. New query function: StringFormatBinary. The function formats an arbitrary value using one of the following formats: - b (default) = <varbinary> -- with the type name
- B = <varbinary, XXX b> -- with the number of bytes
New query function: StringFormatGeom. The function formats a geometry value using one of the following formats: - b = <geom>
- B = <geom, XXX b> -- with the number of bytes
- t (default) = <geom, line> -- with the type of geometry value
- T = <geom, line, 200 c / 3> -- with the number of coordinates and the number of branches after a slash; if the number of branches is 1, it is not shown
The StringFormatGeom query function detects geoms with Z values and reports them as 'areaz', 'linez', 'pointz'. New query function: StringFormatTile. The function formats a tile value using one of the following formats: - b = <tile>
- B = <tile, XXX b> -- with the number of bytes
- t (default) = <tile, 100 x 100 x 3> -- with the width and height and the number of channels; if the number of channels is 1, it is not shown
- T = <tile, 100 x 100, float64x3> -- with the width and height and the pixel type that includes the number of channels
New query function: StringFormat. The function formats an arbitrary value according to its type. (GEOMWKB values are currently reported as binary to avoid conversion, similarly to previous builds. We will change this to report GEOMWKB values as geometry in the future. Also, the number of bytes reported in the 'B' format by StringFormat / StringFormatBinary may differ from the number of bytes reported in the 'B' format by StringFormatGeom / StringFormatTile. The former two functions report the number of bytes in the value as it is stored in a MAP file, with some compression already applied. The latter two functions report the number of bytes in the same value expanded to RAM, with no compression.) The StringFormatNumber query function now uses language-specific decimal separator for the default format. The StringFormatVector query function now uses language-specific list separator for the default format. (Fix) The GDB dataport marks computed fields for length and area so that they do not prevent copy and paste of tables into a MAP and get converted to non-computed fields instead. The Clip transform automatically removes records with NULL geometry when producing a new table. (Fix) The LAS dataport no longer sometimes returns incomplete data for a spatial search. Error reports produced by the GDB dataport and dataports for databases like SQL Server are extended to cite the names of relevant fields and indexes. End of list.
|