Online Analytical Processing is a technology used to organize business databases and allow the execution of multidimensional analysis of corporate data. It supports multiple analyses by different users, as well as all possible browsers. It allows the user to select all information and analyze segmented users, while simultaneously decreasing the quantity of reported data.
This customization can be seen in the multidimensional view structure, choosing fields that select the level of aggregation and dimension and/or a specific set of data. It allows the user to rotate, drop attributes, browse, and expand or collapse displayed data.
The user can access this tool by clicking on:
1 About the OLAP Tool
Axional OLAP (Online Analytical Processing) is a tool aimed at improving the agility of queries for large quantities of data, showing the results quickly and intuitively in table or graph form.
Designed to handle large volumes of data and convert them into useful information, with OLAP users gain a big-picture view of the company’s position, thus facilitating effective decision-making.
The key feature of OLAP tools is the rapidity with which they return responses. This is possible thanks to the application’s use of multidimensional structures, or OLAP Cubes, which contain summarized data from large databases. This concept is used as an analytical tool in all areas of business, such as sales, marketing, executive reports, etc.
1.1 OLAP Cubes
An OLAP cube is a multidimensional data structure, in which physical data storage is carried out in a multidimensional vector. OLAP cubes can be considered an extension of the two dimensions present in a spreadsheet.
To this end, data is laid out in vectors which facilitate rapid analysis. These vectors are what we call cubes.
To provide an example, a company could analyze certain financial data by product, by period, by city, by type of revenue or expense, or by comparison of actual data vs. budgeted figures. These parameters, the basis on which data is analyzed, are known as dimensions.
Data can be accessed simply by indexing based on the values of dimensions, or axes.
A financial analyst wants to view data in various forms, e.g. a view based on all cities (which could appear on the x-axis) and all products (on the y-axis). This data may correspond to a specific period, version, or type of expenses. After viewing data in this specific layout, the analyst may then wish to view data from a different perspective, and be able to modify it immediately.
The cube can adopt a new orientation, allowing data to now appear according to period or type of expense. As this reorientation entails the need to summarize a very large quantity of data, this new view must be generated efficiently so as not to waste the analyst’s time. In other words, the process must occur in seconds, rather than the hours that would be necessary in a conventional relational database.
The set of Dimensions of Measures is called a Cube. Cubes facilitate multidimensional analysis in order to respond to companies’ complex informational needs. Cubes are constructed based on the XML standards of Mondrian.
A business-based point of view, useful for data analysis through what is often called a hierarchy. Examples include time, product, or geographic location.
A dimension may contain one or more hierarchies. Within each hierarchy, the order of some or all levels is specified, in a dimension used to determine routes for navigation. Example: Country, Region, Province, City, Neighborhood.
Dimensions often contain multiple levels of detail. For example, the time dimension may be comprised of Year, Quarter, and Month levels. For dimensions with only one level, the level is treated as implicit. As such, when referring to the physical representation of data, a dimension’s level is often simply called a dimension.
A given value for a level of a dimension. For example, the time dimension’s Year level could have members such as 1998, 1999 and 2000. The number of unique members in a given level of a dimension is referred to as the cardinality of that level.
A fundamental, aggregated measure of business, such as sales or profits. Measures also have statistics associated with them, such as Total, Unit Count, and Average.
1.2 Dimensions and Hierarchies
Each dimension of an OLAP cube can be summarized by a hierarchy. For example, the time scale (or dimension) “May 2013” may be included within “Second Quarter of 2013”, in turn included in “Fiscal Year 2013”.
Likewise, in cube dimensions which represent geographic position, cities may be included in regions, nations, or continents. Products may be classified by category, and expense budget lines may be grouped by type of expense.
The analyst may begin on a highly summarized level, such as the total difference between actual results and budgeted amounts. Subsequently, they may descend through the cube and its hierarchies to observe the greatest level of detail possible, allowing them to discover the places in the cube responsible for the difference, according to product and accounting period.
All dimensions, whether linear or hierarchical, allow the user to define a set of properties or attributes which contain additional information about the dimension’s members. For example, for the dimension “Store Name”, information such as store type, manager name, and address may be included.
Each element of a dimension is a member of that dimension. For example, February 2013, March 2013, Quarter 1 2013 and Fiscal Year 2013 may all be members of the dimension Time.
Dimensions may be linear or hierarchical. Examples of hierarchical dimensions could include time, brands, sales, products or clients. Their structure could appear as follows:
There are two possible types of hierarchies:
- Parent-child: corresponds to the dimensions created in the same reference tables.
- Multilevel: defined by the fields of one or more tables, logically linked together by the relationship between one or more of the fields.
Members of a dimension can be related to each other in a hierarchical way. For example, a specific day belongs to a specific month which, in turn, exists within a specific year. To reflect these relationships, members of the dimension are organized into dimensional hierarchies.
A dimensional hierarchy is a logical structure which utilizes ordered levels as a means of data organization and aggregation. For example, the Time dimension may contain a hierarchy to aggregate data from the Month level into the Quarter and Year levels.
Axional OLAP supports cubes with an unlimited number of dimensions, thus allowing analysts to view data from various angles or “pivots”.
Measures are quantitative data which can be evaluated or measured, and which are available in source databases. For example, sales volume or invoice amounts are typical cube measures. Measures represent data which can be examined and analyzed in matrices and graphs. Possible examples include Sales, Costs, Margins, etc.
In general, a measure is a data type whose information is used by analysts (users) to measure the performance and behaviour of a business process or object, such as quantities, sizes, sums, durations, etc. Prospective measures which may be included in a cube are the same numeric values used in summary calculations.
Cubes may also contain measures calculated from arithmetic operations performed on tables’ individual measures. The system allows the definition of very complex calculated measures, such as, for example, comparison of a given month’s sales with sales from the same month in previous years.
Measures have dimensions which classify their data. For example, the measure Sales may have Product, Time, and Location as dimensions. When a measure has a concrete dimension, we say that the measure is “dimensioned” by it. For example, Sales is dimensioned by Product. The group of dimensions for a measure constitutes the dimensionality of that measure. For example, the dimensionality of Sales is Product, Time, and Location.
2 Getting Started
2.1 Using an Existing Database
Typically, the definition of cubes, as well as the loading of data, are external processes not handled by users. When using the tool, the user only needs to select the Database with which they will work, the Schema and the Cube.
The dimensions and measures that the user can work with will be shown automatically according to the definition of the selected Cube.
2.1.1 Creating and Populating a New Database; Loading Excel Files
The creation or modification of a Cube Schema is reserved for specialized users with a sufficient level of technical expertise. The creation/modification procedure is described in the section Building a Cube.
3 Using the Tool
3.1 Description of Main Page and Panels
The concepts required for correct use of the OLAP tool, gaining the maximum possible advantage from it, are outlined below. The following image shows the various work panels used to perform queries:
The panels are as follows (see Section 3 for a description of their functions):
- OLAP Data Source
- Data Display
- Additional Options
3.1.1 OLAP Data Source
This section (Panel 1) allows the user to choose the database on which they will perform queries, as well as the schema and the cube.
To set up the tool, the user must choose, in this order:
- The database: set of structured data to which the tool will be applied. Any database (Informix, Oracle, SQL Server, DB2, PostgreSQL, MySQL) accessible to the user may be chosen.
- The schema: functional organization of the database into a set of cubes which allow information analysis. It also contains the rules to create these cubes, oriented towards a specific type of analysis.
- The cube: set of analytical vectors which allows information to be viewed via various multidimensional criteria. It can be considered a representation of existing relationships between data in a database. Cubes are comprised of dimensions and measures.
Once Panel 1 selections have been completed, the tool will automatically fill in the available dimensions and measures for the selected cube.
Within this box (Panel 2), we can find dimensions grouped by hierarchy and levels grouped according to similarity, according to the user’s perception of the data.
- Hierarchies: represented with a green dot; group different levels of the same concept.
- Level: represented with a blue dot; exist within a hierarchy. It is worth noting that the user may encounter properties within a Level. These properties show extra data about the Level in question, but it is important to mention that properties cannot be filtered.
In the following image, we can observe an example of a dimension. The dimension in question is “Store”, and is composed of a hierarchy also called “Store”. This hierarchy is composed of two Levels, “Store State” and “Store Name”. The latter Level also has “Store Manager” as a property. This is a concise and intuitive way to organize and select data to carry out your own analysis.
Panel 3 shows the measures available for queries.
Measures are data which can be measured quantitatively. For example, sales volume or invoicing are measures in our cube, and when they intersect with dimensions such as product or customer, they allow us to obtain measure totals for the values of those dimensions.
Two types of measures exist:
- Direct: values which represent existing data taken directly from the cube’s source database. Represented with a red dot.
- Calculated: values obtained from a calculation involving multiple direct measures, whose form of calculation has been defined in the cube creation schema. Represented with an orange dot.
The dimensions and measures chosen by the user are shown in the central column of the application. This column contains three boxes: one for rows, one for columns, and the third for filters. By default, the dimensions selected are placed in rows and measures in columns; however, they may be organized differently according to the user’s needs. In section 4, Creating a Query explains how the user can change this default order.
In this box (Panel 4), the user places data which they wish to appear in the rows of the resulting table.
Both dimensions and measures can be used in rows. The user may insert as many dimensions as needed for analysis, but should keep in mind that the final result will combine all values of each dimension with all values of the measures. As such, the final result may be difficult to represent and inconvenient to interpret.
Filters may be applied to the levels chosen, e.g. to filter by value, determine an ordering criteria, etc. All of these options are explained in section 4, Creating a Query.
If the user decides to add measures in this panel, these will be shown pivoted in the final result.
In this box (Panel 5), the user places data which they wish to use as columns in the resulting table.
Both dimensions and measures can be used in columns. Each measure represented in the columns of the final table will take a value corresponding to each of the values in its row. For example, if the row contains the level Unit Sales and the column contains the level Product, the final result will be the Unit Sales total for each Product. Levels can also be added to create a column for each contained value. Just as in the previous case, creating too many columns may make the final result difficult to interpret.
Filters can also be applied to columns. If the user decides to add dimensions here, these will be shown pivoted in the final result.
In this box (Panel 6), the user places filters applied to the query which they wish to perform.
Both dimensions and measures can be used in this panel, and the user can insert as many levels as needed. The objective is to filter the final result according to these levels, without the levels themselves appearing in the final table.
3.1.7 Managing Query Toolbar
The right-hand side of the screen contains the toolbar, with the following features:
- Clear Axes: this option allows the user to delete an entire query and restart.
- Execute Query: the user must click this button to run a query.
- Favorites:by selecting this tool, the user can insert a label to name a query and save it in the Favorites tab, taking into account the format. Three possible options exist:
- Export: exports files to XLS, XLSX, CSV or txt format. If the user has added the function "Toggle Clear Groups" (see below), they should keep in mind that when exporting to Excel, they will not be able to sort rows and columns. This is because the repeated values needed to sort/order will be missing.
More (drop-down menu):
- Esport as channel: select this option to include this specific query as a channel on your desktop, recommended for most commonly used queries.
- Toggle clear <br>/groups: function used to hide or reveal repeated fields in the rows/columns of a query.
- Force Refresh: with this flag, the user tells the application to automatically recalculate results after any modification to a query, without the need to click the “Execute Query” button.
- Swap: through this function, the user can swap rows and columns with just one click.
- Small/Big: increase or decrease font size
- Info: external link to the application’s documentation.
3.1.8 Data Display
The results of executing a query are shown in Panel 8. By default, the result is displayed in table form according to the rows and columns selected.
The tool offers various display options, which are described in section 6.
3.1.9 Additional Options
Additional Options provides users with a panel of functions explained in more detail in section 8 Additional Options.
4 Creating a Query
To create a new query, the user must follow these steps:
- If the tool already has a query created, the registry must first be cleared using the toolbar button at the top of the screen (panel 7).
- Next, the user chooses the Dimensions and Measures to be displayed, located in Panels 4 and 5. Important note: both Dimensions and Measures can be placed either in Columns or Rows. However, the same element cannot be repeated in both a column and a row.
The following video shows how to create a query using rows and columns.
- After this step, queries are ready for execution. When the green “play” button in Panel 7 is clicked, the result will be shown (initially in table form) with the rows and columns selected for the query. In Section 6, available viewing and formatting options are described.
Depending on the estimated time needed to return results, a panel may appear indicating that the operation is in progress.
To modify the selection of rows and columns, various options are available:
- Change section. By dragging and dropping, a row or column can be moved from one panel to another (between Panels 4, 5 and 6).
- To remove rows or columns from a panel, the user should click the which appears by hovering the mouse over a panel or right-clicking.
- Swap rows/columns. Using the Swap option in Panel 7 (drop-down menu More), rows and columns are exchanged, meaning that Panel 4’s content swaps with Panel 5’s. As a result, the rows and columns of the resulting query table are also swapped.
4.1 Column Position
In tables generated by queries, the first columns will coincide with the dimensions chosen. In other words, if a user chooses the dimension Item, the first column will be the Item dimension. After that, the columns themselves (specified in Panel 5) will be displayed in the same order as the Panel.
If a dimension is placed in the columns, then measure columns will be duplicated for each value of the dimension. The manner of grouping will depend on the configuration of the “Group pivoted measures” parameter (see Settings section).
4.2 Dimension Hierarchies
As discussed in the introduction, dimensions tend to be organized by dependency relationships called hierarchies. Hierarchical organization is very important, as it conditions the ways users can execute queries and represent data.
When various dimensions from the same hierarchy are chosen, the resulting columns are always shown by order of dependency, not by the placement order in Panel 4.
4.3 Managing the Time Dimension
The dimension Time offers additional hierarchy options.
Normally, separate hierarchies are defined for time elements all contained in the same year, above those not contained in the same year. For example:
- Contained Hierarchies:
- Year – Quarter – Month
- Year – Half-Year – Quarter – Month
- Uncontained Hierarchies (a week can begin in one year and end in another):
- Year – Week – Day
An option called Time-Compare is also available. In this option, the Levels of the hierarchy are defined as unique. Using these dimensions, the result shown compares values from more detailed levels (e.g., month or quarter) with each of the higher levels (e.g., quarter or year).
For example, if the user selects two Time-Compare dimensions (year and quarter) with the measures “Store Cost” and “Store Sales”, the following result will be generated:
As seen in this image, the measures (Store Cost and Store Sales) are represented as a bar chart, dividing the data into years and quarters.
Properties are data only shown in the results table, but not directly involved in the query. Logically, they can only be selected if the Dimension associated with them is included in the query.
If the Dimension is in the Rows panel (Panel 4), then Properties columns will always be placed to the right of their corresponding Dimension.
If the Dimension is in the Columns panel (Panel 5), then the values of Properties will be placed in the column headers.
Query results can be filtered by Dimension as well as by Measure. Filters can be applied on dimensions/measures included in the results displayed (Panels 4 and 5). They can also be applied on dimensions/measures which are not shown in results but which limit the result set (Panel 6).
4.5.1 Filtering Dimensions
- For dimensions with Non-Unique values, the only available option is Filter by Value.
- For Unique dimensions (indicated by a key symbol), the additional option Filter by Pattern is available.
Filter by Value
The tool will open a panel to configure the filter:
The left-hand box shows possible values for the Dimension. The right-hand box will contain the values selected to be filtered. The buttons between the boxes are used to add or remove values from the filter.
By default, the filter type is set to Inclusive, meaning that only values selected in the right-hand box will be shown in results. To perform the reverse operation, Exclusive filtering, click the slider button below the right-hand box. This will make the query results show all data except those indicated on the right.
When the lists of values in the boxes are extremely long, there will be a field at the top of each box to search for and filter specific values. This function searches only for values which contain text entered in the search field.
To optimize response time, if the list of values is extremely long, the left-hand box will only show the first 1,000 values. In these cases, an initial filter can be set in the field at the top of the form. This will return a subset of desired data to configure the query filter.
Filter by Pattern:
This option is used to filter by similarity. In other words, a specific “house” value will be compared with the pattern specified.
Requesting this type of filter shows a form with filter configuration options. In the case of dimensions with textual values, the options form will look like this:
First, the Mode of Comparison must be indicated. The following comparison operators are available:
- Equal to
- Not equal to
- Does not contain
- Starts with
- Ends with
- Less than or equal to
- Greater than or equal to
In the left-hand field, we will place the text to be compared. If the operator “Between” is chosen, then both fields will be activated in order to specify the two values to be compared.
In the case of dimensions with numerical values (including dates), the options form will look like this:
First, the Mode of Comparison must be indicated. In this case, only the following comparison operators are available:
- Equal to
- Not equal to
- Less than or equal to
- Greater than or equal to
In the left-hand field, we will place the value to be compared. If the operator “Between” is chosen, then both fields will be activated in order to specify the two values to be compared.
If the user filters by more than one Level belonging to the same hierarchy, the application will remember this order. When the Filter by Value form for a Level is shown, previously-set filters will be applied when selecting available values.
For example, if a user filtered the Country Level by the value “Spain”, then later filtered cities, only Spanish cities would be shown and non-Spanish cities would be discarded.
4.5.2 Filtering Measures
First, users must also select the Dimension (Row in Panel 4) to which the filter will apply. This type of filter is very powerful and can be configured to perform a wide range of operations. Once the Dimension has been chosen, two tabs will appear:
Filter by Best/Worst (Ranked or Top 10):
The first tab is used to set filters by rank (ranked results).
The first step is to choose the type of rank: First or Last. The right-hand field indicates the number of results to retrieve. This type of filter is used to create queries such as, for example, “Find the 10 stores with the best sales”.
Filter by numerical comparison:
The right-hand tab specifies which traditional comparison filters apply. The tool allows users to simultaneously set multiple filter conditions. To add a condition, press the green “+” button.
As shown in this image, for each filter users choose a comparison operator from the left-hand menu and a numerical value to be compared in the right-hand field.
The available ordering options are:
- Automatic: The default option. The OLAP tool decides the best order according to the hierarchies selected.
- Force no sorting: Instructs the tool not to apply any ordering criteria derived from hierarchies for the given column.
- Sort ascending: The results table will be sorted in ascending order for the given column.
- Sort descending: The results table will be sorted in descending order for the given column.
Because of the hierarchies dependencies, users should keep the following rules for ordering options in mind:
- Only one Measure can be chosen for ordering.
- Ordering grouped data rule: when a table with grouped results is ordered by a Measure, the order is applied internally for each value in the main group.
It is set to sort by descending order according to sales. The result would be the following:
Sales are sorted in descending order for each individual store.
If the user wants to sort all tables, ignoring grouping, they must use the Group/Ungroup option in Panel 7. Next, they must deactivate the dimension’s ordering using the Force No Sorting option.
The query result would appear as follows:
- To clear ordering/sorting, simply change the setting to Automatic.
4.7 Grouping and Subtotals
The OLAP tool automatically groups Dimension of the same hierarchies. If the user wants to establish subtotals, they will be defined by these groups.
To create the subtotals the user must set rows and columns as described below.
To define rows for Subtotals, this procedure must be followed:
- First, the user chooses the Measures (columns) which they wish to aggregate. The Subtotal Types option opens a form where the user can indicate the type of calculation they wish to perform:
The most frequent type of calculation is Sum, but Average, Minimum and Maximum are also available. Multiple options can be chosen simultaneously. The tool will display a row for each calculation option.
- Second, the user chooses the Dimension (row) which will determine the breakpoint for the calculation. Multiple dimensions can be chosen, and the number of subtotals generated will equal the number of rows selected.
The following query has Store and Quarter selected as its two breakpoints. The subtotal (sum) of the Sales column should appear as follows:
The result will look like this:
To define the columns of Groups/Subtotals, this procedure must be followed:
- First, users must ensure that at least one Dimension is specified in the Columns panel (Panel 5).
- To aggregate a Measure horizontally, users must choose the Subtotal Types option. This will open a form to select the type of calculation to be performed. Various types of calculation can be chosen simultaneously.
- In addition to the column, the results table will also show a final row of subtotals.
For example, the following query is designed to return cost subtotals of various product families by store and by quarter:
The results table will have a row of cost subtotals by product family:
IMPORTANT: only if user has enabled the Total by column option in the Settings menu (see Settings section), the result table will show an extra column with the total Cost by quarter. See example below:
If the user has added the function "Toggle Clear Groups", they should keep in mind that when exporting to Excel, they will not be able to sort rows and columns. This is because the repeated values needed to sort/order will be missing. The result will look like this:
4.8 Percentage Columns
For Measures, an additional column can be added to the query which shows the percentage of the total which said column represents.
4.9 Add dimension as an aggregated measure
Although the dimensions are discrete values (limited to a set), they can be transformed and used as quantities over another measure. The example shown below has one dimension (Story City) and one measure (Unit Sales). Right click on a new dimension to be aggregated (Promotion name).
Select desired aggregation option by double clicking on it. The agrregation options are:
- Count: counts number of the selected dimension in relation to the measure.
- Distinct count: counts number of different selected dimension in relation to the measure.
- Min: counts maximum number of selected dimension in relation to the measure.
- Max: counts maximum number of selected dimension in relation to the measure.
In this example it is shown Story City and the unit sales. The aggregated COUNT Promotion Name counts the number of promotions in relation to the unit sales.
TO DOThis section is incomplete and will be concluded as soon as possible.
5 Drill Down / Roll Up
Option to drill down into the data from the table.
To access this option left-click the row and click the window "drill down" which appears.
The tool will open a panel to configure the drill down options.
The definition of the parameters is like creating a query.
Mark the desired options and click "ok" button to generate the results of the drill down. To modify the parameters click "back" button.
If you do not remember the selected row, right-click on the information icon.
6 Display Results
Axional OLAP allows results to be viewed both in table form and as various types of graphs and charts. Results may also be displayed as maps if the data contains a geographic component.
The various modes of representing results offer users an easy way to gain new perspectives on their data, thanks to a highly interactive and user-friendly interface. The formats used to represent results are chosen in the toolbar of the Data Display Panel (Panel 8).
Located in the upper right part of the data panel, the graphic properties of the query results can be changed:
- Data: the interface displays the data in a table (by default).
- Chart: the data is display as a chart. Two new selectable options will be displayed: select chart type and chart stocking options.
- Map: this option is only available when geographical dimensions have been selected.
- Info: shows the query like a SQL statemen or XSQL-Script.
6.1 Table Format
The default option is the (Data) Table format. Upon executing a query (using the green “play” button in Panel 7), the result is initially shown as a table with rows and columns selected for the query.
The data shown in the table’s cells are formatted according to their type (text or numerical). The tool offers various options for the user to modify the data presentation format:
- Cell graphics. Instead of numeric data, a column’s cells can display miniature graphics inside. To choose this type of view, click the right-hand button above the column header or click this icon
It should be noted that the two colors available for data are displayed depending on the selection column range. Some of the available mini-graphics include:
Dots: the cell contains a dot whose diameter corresponds to the cell’s relative value, compared to the maximum value in the column.
Maximum/Minimum: a horizontal line is displayed on either side of a vertical line which marks the center value, with an icon (dot) on the opposite end. If positive and negative values coexist, the dots are placed on opposite ends of the central vertical line, with different colors.
Percentage Bar: displays a horizontal bar whose length corresponds to the cell value. IMPORTANT: this option is only supported for columns that show percentages.
The length of the bar corresponds to its value, not factoring in the sign (positive or negative). The bar displays in different colors for positive and negative values.
- Dots: the cell contains a dot whose diameter corresponds to the cell’s relative value, compared to the maximum value in the column.
Conditional Format: for columns with numerical values, the user can set negative values to be shown in red, or vice versa.
Column Width: by dragging the right-hand border of the column with the mouse, width can be adjusted.
6.2 Chart Format
After selecting this option in Panel 8, the user can choose between different types of graphs and charts. Depending on the type of graphing, Panel 8 will show a selector to choose specific viewing options.
- Column: Column values (measures) are shown as vertical bars. For each column, the number of bars matches the number of measures. Each bar has a distinct color. This type of chart offers the following options:
Bar: The values of rows (measures) are shown as horizontal bars. The display of the bars depends upon the number of columns (equal to its number of measures). Just as in the previous case, the color of each bar is different and also offers the options to standardize, stack, and stack 100%, depending on the user’s needs.
Area: Represents data from rows as a line chart. The area between the line and the axis is shaded in with a certain color. This type of chart can also be represented as standardized, stacked, and 100% stacked.
Line: Row values are shown as lines. For example, if two measures exist, two lines will be displayed. This will also vary according to the number of columns.
Circular (Pie): A circular chart used to represent proportions and percentages. Each proportion corresponds to a measure.
Dispersion (scatter plot): This type of plot represents only the inflection points of the dimensions chosen.
Treemap: This type of chart provides a hierarchical view of data. As with the pie chart type, the Treemap is divided according to the number of columns (measures).
Heat: This type is used to visualize the relationship between two variables that belong to different categories. The value is determined by color gradation.
Not all types of graphs and charts make sense for a given query result. Although the OLAP tool can construct any chart selected, the user should make sure to choose a type adapted to the nature of their query result. For example, a bar chart is well-suited to comparisons of current values with values from previous years, while a pie chart would not be appropriate.
6.3 Map Format
Axional OLAP includes geographical analysis capabilities. In general, a great deal of analyzed data has a component referring to a place or position, e.g. an address, postal code, country, etc.
By choosing the Map Format option, geospatial representation allows users to visualize and analyze data with a geographic component, helping them discover patterns and trends that would remain hidden in table form. These maps can indicate places with points or heats.
This is an example of a map with indicative points.
This is an example of a map with indicative heats. This option allows the users to specify which information desires to see of the selected measures.
7 Generating Reports
Not available in the current version.
8 Additional Options
It will appear the following panel:
8.1 OLAP Studio
By selecting this option, the user will be redirected to the main screen of the OLAP tool where queries are created.
Shows the selected Schema’s XML code.
Table which contains the last queries of the users and their statistical details. It appears search information such as number of times of each query, performance time, number of rows and number of columns. By left-clicking on the rows, users can see the reference, the dimensions, the measures and the filters applied to the query below the table.
Lists of errors and alerts produced during the Cube processing.
The Accelerator is a product designed to help users improve query performance. This tool processes queries more quickly than the traditional database server, and improves performance while simultaneously reducing administration work. A database accelerator is particularly useful with an OLAP application because its queries typically involve large amounts of data. For example, OLAP schemas can help make decisions based on company-wide consumer trends over an extended time period, or analyze sensor measurements (electricity consumption of end users), etc. Query times can be reduced to 10% or more of their typical time by using a database accelerator.
To use this tool, a database which has an accelerator available must first be selected.
If the user experiences an error due to the accelerator, or simply wants to inform themselves, they can view all accelerator information in this panel. The panel is divided into three sections:
Some of the most useful metrics include “Average Query Queue Wait Time”, showing the user the average wait time for a given query, as well as “Number of failed query requests due to invalid state”, which shows the user the number of errors for a given query.
This corresponds to the data marts within the accelerator. This data is defined in wic_conf. For example, data mart “tpch_1” tells the user when the retrieved data dates from.
This corresponds to the data marts within the accelerator. This data is defined in wic_conf. For example, data mart “tpch_1” tells the user when the retrieved data dates from.
Tables and columns included in the data mart, as well as the relationships between them. This panel helps users in situations where errors are produced when executing a query. The panel will show that the dimension or measure selected is not included in the table, thus letting the user know the source of the error.
Allows the user to choose whether confirmation should be required for large result sets, and whether alerts should be generated when a chart has been trimmed. Users can also define table characteristics (hide rows/columns, set column size, etc.), and choose the type of execution (“automatic” or “hide empty values”).
This part of the menu is used to set configuration or general parameters, which will affect how the OLAP tool behaves and how it presents results.
Don’t Ask Confirmation for Potentially Large Result Sets: Some queries may return large volumes of data.
As a result, performance may be affected. In many cases, Axional OLAP can predict this issue and will ask the user
for confirmation if a query’s result set exceeds a certain size. The user should check the “Don’t Ask Confirmation for Potentially Large Result Sets” box to avoid these prompts.
Alert: Result Set Used in Chart Has Been Trimmed: This alert notification appears when the result set is so large that it must
be trimmed to be shown on a chart. As an example, imagine that a user executes a table with 2 million clients.
The following notification would be shown to warn them that results will be separated.
- Hide Rows/Columns: option used to reveal or hide rows and columns with null values or zero values. As an example, imagine that a user wanted to create a query about clients who consumed their product in the past 12 months. A table would be executed which would hide or reveal null or zero values, according to how the user set this option.
- Total by columns: Whenever possible, when performing subgrups, the total will be added not only in the last row (showing columns total), but also in a new column showing row total.
8.7 Query History
- Last Queries: All queries executed in the OLAP system are stored and recorded so that users can recover them at any time.
- Group Favorites: files shared by a group as Favorites will be saved in this location.
- Favorites: queries executed and saved as favorites.
- Running: list of executions in progress that have not yet reached the server executions.