When working with numbers and applications there are limitations on how precise the application can return these numbers. The number of significant digits that can be accurately represented by a computer program depends on the data type being used.
In the case of TrendMiner, the most precise data type available can only show up to a certain number of significant digits. This means that if you enter a number with more digits, TrendMiner may not be able to represent it accurately.
TrendMiner uses different data types between hubs, and thus the significant numbers in TrendMiner differ as followed:
- TrendHub: 7 significant digits
- ContextHub: 15 significant digits
Other than the significant digits there are also limitations on the supported ranges.
The supported ranges for each hub are:
- TrendHub: -1e38 to 1e+38
- ContextHub -1.79E+308 to 1.79E+308
Significant numbers and input fields
TrendMiner aims to provide transparency to the user regarding the numerical inputs that are being used in computations behind the screens. To achieve this input fields within TrendMiner have been implemented with the following features that will automatically adjust the user's input:
- When entering values with more significant digits than the maximum supported, we will automatically adjust the input to the maximum supported significant digits.
- When entering values outside the supported range, we will adjust the input to the nearest allowed value or return a message that the value is out of range.
To support small values, we allow the display and input of values in scientific notation for any input field. To enter a value as scientific notation, you can use the E-notation.
Significant numbers and tag values
Default behavior
TrendMiner has default settings for displaying numerical values based on their magnitude. The default behavior is as follows:
- For values ABOVE 1, we display the values in decimal notation with one decimal place (0.1 accuracy) unless the value is greater than or equal to 1E6, in which case we use scientific notation. Please note that for very large values that cannot be displayed accurately, we will show as many digits as possible. For example, 1000234 will be displayed as 1.00023E6.
- For values between 1 and 0.001 (inclusive), we display the values in decimal notation with three decimal places. Examples:
- 0.001234567 will be displayed as 0.001
- 0.1234567 will be displayed as 0.123
- For values below 0.001, we display the values in scientific notation with two significant numbers.
- 0.0001234567 will be displayed as 1.2e-4
- 0.00001234567 will be displayed as 1.2e-5
- 0.000001234567 will be displayed as 1.2-6
Custom settings
TrendMiner also allows users to customize the display of values based on the accuracy setting. This setting can be found in the tag details menu.
Note: In some cases, we may not be able to use the configured accuracy due to the limitations of significant numbers. In these cases, we will display the maximum allowed significant numbers.
- For values above or equal to 1E+06, scientific notation is used, and the accuracy setting determines the number of significant digits shown.
Examples are provided in the following table:
Accuracy | Input | Result |
100 | 8123456 | 8.1234E6 |
1000 | 8123456 | 8.123E6 |
0.1 | 8123456 | 8.123456E6 |
0.1 | 81234567 | 8.123456E7 |
- For values between 1E+06 and 1E-03 (inclusive), decimal notation is used, and the accuracy setting determines the number of significant digits.
Examples are provided in the following table:
Accuracy | Input | Result |
100 | 812345 | 812300 (incorrect configuration by user) |
0.1 | 123.123 | 123.1 |
0.01 | 123.123 | 123.12 |
0.01 | 0.0123 | 0.01 |
0.01 | 0.1 | 0.10 |
0.001 | 0.00123 | 0.001 |
1e-18 | 0.0123456789 | 0.01234567 |
- For values below 1E-03, scientific notation is used, and the accuracy setting determines the number of significant digits.
Examples are provided in the following table:
Accuracy | Input | Result |
0.1 | 0.00001 | 0 (incorrect configuration by user) |
1e-4 | 0.00010 | 1E-4 |
1e-4 | 0.00012 | 1E-4 |
1e-6 | 0.0001 | 1.00E-4 |
1e-18 | 0.0008123456789 | 8.123456E-4 |