Sökfrågeutformning

I Designvyn för sökfrågor kan du skapa och redigera en databassökning.

De flesta databaser använder frågor för att filtrera eller sortera databastabeller och visa poster. Vyer fungerar på liknande sätt som frågor, men finns på servern. Om databasen finns på en server som stöder vyer, kan du använda vyerna om du vill filtrera posterna på servern så att de visas snabbare.

Om du markerar kommandot Skapa vy på fliken Tabeller i databasdokumentet ser du fönstret Vyutkast som påminner om Sökfrågeutkast som beskrivs här.

Layouten för fönstret Sökfrågeutkast sparas med en skapad sökning men kan inte sparas med en skapad vy.

Vyutkast

Om du vill skapa en sökfråga klickar du på ikonen Sökning i ett databasdokument och klickar sedan på Utkast till databasformulär.

Den nedre panelen i Designvyn är där du anger sökfrågan. För att skapa en fråga, ange databasfältens namn som du vill inkludera och villkoren för att visa fälten. För att byta ordning på kolumnerna i den nedre panelen i designvyn drar du kolumnrubriken till den nya platsen, eller så väljer du kolumnen och trycker +pil-tangent.

In den övre delen av designvyfönstret, ligger listen med ikonerna för Frågeutformningen samt Design-listen.

Om du vill testa en sökfråga dubbelklickar du på sökningens namn i databasdokumentet. Sökningens resultat visas i en tabell. Obs! Tabellen är temporär och sparas inte.

Tangenter i Sökfrågeutkastläge

Bläddra

När du öppnar sökfrågeutformaren för första gången visas en dialogruta där du först måste välja den tabell eller fråga som ska ligga till grund för din nya fråga.

Dubbelklicka på ett fält om du vill lägga till det i sökningen. Dra och släpp för att definiera relationer.

När du utformar en sökfråga så kan uttrycket inte användas för att ändra de valda tabellerna.

Ta bort tabeller

För att ta bort en tabell från sökfrågeutkastvyn klickar du på tabellfönstrets övre kant och öppnar snabbmenyn. Med kommandot Radera tar du bort tabellen från vyn. Du kan också använda tangenten Delete.

Flytta tabeller och ändra tabellstorlek

Du kan flytta runt tabellerna som du vill och även ändra deras storlek. När du ska flytta en tabell tar du tag i dess övre kant med musen och drar den till önskad position. För att förstora eller förminska storleken drar du i kanten eller ett hörn.

Relationer mellan tabeller

Om det finns datarelationer mellan ett datafält i en tabell och ett datafält i en annan tabell kan du använda de här relationerna för din sökning.

Om du till exempel har en tabell för artiklar som identifieras med ett artikelnummer och en tabell för kunder där du registrerar alla artiklar som en kund beställer med motsvarande artikelnummer, finns det ett samband mellan de två datafälten ”artikelnummer”. Om du nu vill skapa en fråga som returnerar alla artiklar som en kund har beställt måste du hämta data från två tabeller. Denna relation mellan data i de två tabellerna måste du uttrycka explicit för LibreOfficeDev för att det ska kunna ske.

För detta, klicka på ett fältnamn i en tabell (till exempel fältnamnet Artikelnummer i tabellen Kund), håll musknappen intryckt och dra sedan fältnamnet till fältnamnet i den andra tabellen (Artikelnummer i tabellen Artikel). När du släpper musknappen visas en linje som förbinder de två fälten mellan tabellfönstren. Det villkor som gäller, att innehållet i de båda fältnamnen måste vara identiskt, infogas i den resulterande SQL-frågan.

Om du ska skapa en sökning som bygger på flera olika relaterade tabeller måste du använda LibreOfficeDev som gränssnitt mot en relationsdatabas.

Det går inte att använda tabeller från olika databaser i en sökning. Sökfrågor med flera olika tabeller kan bara skapas i en och samma databas.

Ange relationstypen

Om du dubbelklickar på kopplingslinjen mellan två länkade fält eller väljer Infoga – Ny relation … i menyn kan du ange typen av relation i dialogrutan Relationer.

Alternativt trycker du på Tabb-tangenten tills raden är markerad, och trycker sedan på Skift+F10 för att visa snabbmenyn och där välja kommandot Redigera. Vissa databaser har endast stöd för en deluppsättning av de möjliga typerna av sammanslagning (join).

Radera relationer

Om du vill radera en relation mellan två tabeller klickar du med musen på förbindelselinjen och trycker sedan på tangenten Delete.

Du kan också markera respektive poster under inkluderade fält i dialogrutan Relationer. Du kan också trycka på Tabb-tangenten tills förbindelselinjen är markerad, trycka på Skift+F10 för att öppna snabbmenyn och välja kommandot Radera.

Defining the query

Select conditions to define the query. Each column of the design table accepts a data field for the query. The conditions in one row are linked with a Boolean AND.

Specifying field names

First, select all field names from the tables that you want to add to the query. You can do this either by drag-and-drop or by double-clicking a field name in the table window. With the drag-and-drop method, use the mouse to drag a field name from the table window into the lower area of the query design window. As you do this, you can decide which column in the query design window will receive the selected field. A field name can also be selected by double-clicking. It will then be added to the next free column in the query design window.

Deleting field names

To remove a field name from the query, click the column header of the field and choose the Delete command on the context menu for the column.

Saving the query

Use the Save icon on the Standard toolbar to save the query. You will see a dialog that asks you to enter a name for the query. If the database supports schemas, you can also enter a schema name.

Schema

Enter the name of the schema that is assigned to the query or table view.

Query name or table view name

Enter the name of the query or table view.

Filtering data

To filter data for the query, set the desired criteria in the lower area of the query design window. The following options are available:

Field

Enter the name of the data field that is referred to in the Query. All settings made in the filter option rows refer to this field. If you activate a cell here with a mouse click you'll see an arrow button, which enables you to select a field. The "Table name.*" option selects all data fields with the effect that the specified criteria will be applied to all table fields.

Alias

Specifies an alias. This alias will be listed in the query instead of the field name. This makes it possible to use user-defined column labels. For example, if the data field is named PtNo and, instead of that name, you would like to have PartNum appear in the query, enter PartNum as the alias.

In a SQL statement, aliases are defined as follows:

SELECT column AS alias FROM table.

For example:

SELECT "PtNo" AS "PartNum" FROM "Parts"

Table

The corresponding database table of the selected data field is listed here. If you activate this cell with a mouse click, an arrow will appear which enables you to select a different table for the current query.

Sort

If you click on this cell, you can choose a sort option: ascending, descending and unsorted. Text fields will be sorted alphabetically and numerical fields numerically. For most databases, administrators can set the sorting options at the database level.

Visible

If you mark the Visible property for a data field, that field will be visibly displayed in the resulting query. If you are only using a data field to formulate a condition or make a calculation, you do not necessarily need to display it.

Criteria

Specifies a first criteria by which the content of the data field is to be filtered.

or

Here you can enter one additional filter criterion for each line. Multiple criteria in a single column will be interpreted as boolean OR.

You can also use the context menu of the line headers in the lower area of the query design window to insert a filter based on a function:

Functions

The functions which are available here depend on those provided by the database engine.

If you are working with the embedded HSQL database, the list box in the Function row offers you the following options:

You can also enter function calls directly into the SQL statement. The syntax is:

SELECT FUNCTION(column) FROM table.

For example, the function call in SQL for calculating a sum is:

SELECT SUM("Price") FROM "Article".

Except for the Group function, the above functions are called Aggregate functions. These are functions that calculate data to create summaries from the results. Additional functions that are not listed in the list box might be also possible. These depend on the specific database engine in use and on the current functionality provided by the Base driver used to connect to that database engine.

To use other functions not listed in the list box, you must enter them manually under Field.

You can also assign aliases to function calls. If you do not want to display the query string in the column header, enter a desired substitute name under Alias.

The corresponding function in an SQL statement is:

SELECT FUNCTION() AS alias FROM table

Example:

SELECT COUNT(*) AS count FROM "Item"

If you run such a function, you cannot insert any additional columns for the query other than as an argument in a "Group" function.

Examples

In the following example, a query is run through two tables: an "Item" table with the "Item_No" field and a "Suppliers" table with the "Supplier_Name" field. In addition, both tables have a common field name "Supplier_No."

The following steps are required to create a query containing all suppliers who deliver more than three items.

1. Insert the "Item" and "Suppliers" tables into the query design.

2. Link the "Supplier_No" fields of the two tables if there is not already a relation of this type.

3. Double-click on the "Item_No" field from the "Item" table. Display the Function line using the context menu and select the Count function.

4. Enter >3 as a criterion and disable the Visible field.

5. Double-click the "Supplier_Name" field in the "Suppliers" table and choose the Group function.

6. Run the query.

If the "price" (for the individual price of an article) and "Supplier_No" (for the supplier of the article) fields exist in the "Item" table, you can obtain the average price of the item that a supplier provides with the following query:

1. Insert the "Item" table into the query design.

2. Double-click the "Price" and "Supplier_No" fields.

3. Enable the Function line and select the Average function from the "Price" field.

4. You can also enter "Average" in the line for the alias name (without quotation marks).

5. Choose Group for the "Supplier_No" field.

6. Run the query.

The following context menu commands and symbols are available:

Functions

Shows or hides a row for the selection of functions.

Table Name

Shows or hides the row for the table name.

Alias Name

Shows or hides the row for the alias name.

Distinct Values

Retrieves only distinct values from the query. This applies to multiple records that might contain several repeating occurrences of data in the selected fields. If the Distinct Values command is active, you should only see one record in the query (DISTINCT). Otherwise, you will see all records corresponding to the query criteria (ALL).

For example, if the name "Smith" occurs several times in your address database, you can choose the Distinct Values command to specify in the query that the name "Smith" will occur only once.

For a query involving several fields, the combination of values from all fields must be unique so that the result can be formed from a specific record. For example, you have "Smith in Chicago" once in your address book and "Smith in London" twice. With the Distinct Values command, the query will use the two fields "last name" and "city" and return the query result "Smith in Chicago" once and "Smith in London" once.

In SQL, this command corresponds to the DISTINCT predicate.

Limit

Allows you to limit the maximum number of records returned by a query.

If a Limit construction is added, you will get at most as many rows as the number you specify. Otherwise, you will see all records corresponding to the query criteria.

Formulating filter conditions

When formulating filter conditions, various operators and commands are available to you. Apart from the relational operators, there are SQL-specific commands that query the content of database fields. If you use these commands in the LibreOfficeDev syntax, LibreOfficeDev automatically converts these into the corresponding SQL syntax via an internal parser. You can also enter the SQL command directly and bypass the internal parser. The following tables give an overview of the operators and commands:

Examples

Like Escape Sequence: {escape 'escape-character'}

SELECT * FROM Item WHERE ItemName LIKE 'The *%' {escape '*'}

The example will give you all of the entries where the item name begins with 'The *'. This means that you can also search for characters that would otherwise be interpreted as placeholders, such as *, ?, _, % or the period.

Outer Join Escape Sequence: {oj outer-join}

SELECT Article.* FROM {oj item LEFT OUTER JOIN orders ON item.no=orders.ANR}

Querying text fields

To query the content of a text field, you must put the expression between single quotes. The distinction between uppercase and lowercase letters depends on the database in use. LIKE, by definition, is case-sensitive (though some databases don't interpret this strictly).

Querying date fields

Date fields are represented as #Date# to clearly identify them as dates. Date, time and date/time constants (literals) used in conditions can be of either the SQL Escape Syntax type, or default SQL2 syntax.

SELECT {d '1999-12-31'} FROM world.years

SELECT * FROM mytable WHERE years='1999-12-31'

All date expressions (date literals) must be enclosed with single quotation marks. (Consult the reference for the particular database and connector you are using for more details.)

Querying Yes/No fields

To query Yes/No fields, use the following syntax for dBASE tables:

The syntax depends on the database system used. You should also note that Yes/No fields can be defined differently (only 2 states instead of 3).

Parameter queries

Parameter queries allow the user to input values at run-time. These values are used within the criteria for selecting the records to be displayed. Each such value has a parameter name associated with it, which is used to prompt the user when the query is run.

Parameter names are preceded by a colon in both the Design and SQL views of a query. This can be used wherever a value can appear. If the same value is to appear more than once in the query, the same parameter name is used.

In the simplest case, where the user enters a value which is matched for equality, the parameter name with its preceding colon is simply entered in the Criterion row. In SQL mode this should be typed as WHERE "Field" = :Parameter_name

Parameter names may not contain any of the characters <space>`!"$%^*()+={}[]@'~#<>?/,. They may not be the same as field names or SQL reserved words. They may be the same as aliases.

A useful construction for selecting records based on parts of a text field's content is to add a hidden column with "LIKE '%' || :Part_of_field || '%'" as the criterion. This will select records with an exact match. If a case-insensitive test is wanted, one solution is to use LOWER (Field_Name) as the field and LIKE LOWER ( '%' || :Part_of_field || '%' ) as the criterion. Note that the spaces in the criterion are important; if they are left out the SQL parser interprets the entire criterion as a string to be matched. In SQL mode this should be typed as LOWER ( "Field_Name" ) LIKE LOWER ( '%' || :Part_of_field || '%' ).

Parameter queries may be used as the data source for subforms, to allow the user to restrict the displayed records.

Parameter Input

The Parameter Input dialog asks the user to enter the parameter values. Enter a value for each query parameter and confirm by clicking OK or typing Enter.

The values entered by the user may consist of any characters which are allowable for the SQL for the relevant criterion; this may depend on the underlying database system.

The user can use the SQL wild-card characters "%" (arbitrary string) or "_" (arbitrary single character) as part of the value to retrieve records with more complex criteria.

SQL Mode

SQL stands for "Structured Query Language" and describes instructions for updating and administering relational databases.

In LibreOfficeDev you do not need any knowledge of SQL for most queries, since you do not have to enter the SQL code. If you create a query in the query designer, LibreOfficeDev automatically converts your instructions into the corresponding SQL syntax. If, with the help of the Switch Design View On/Off button, you change to the SQL view, you can see the SQL commands for a query that has already been created.

You can formulate your query directly in SQL code. Note, however, that the special syntax is dependent upon the database system that you use.

If you enter the SQL code manually, you can create SQL-specific queries that are not supported by the graphical interface in the Query designer. These queries must be executed in native SQL mode.

By clicking the Run SQL command directly icon in the SQL view, you can formulate a query that is not processed by LibreOfficeDev and sent directly to the database engine.