JMP allows users to specify user-defined formats. These custom formats are available wherever natively-supported formats are available: Data tables, Graph Builder, CSV Import, Data Filters, JSL, etc. These custom formats can be used both for inputting and outputting dates and times.
To format a column of numbers and obtain a numeric result in the data table, you either must use the Column Property dialog and change the format by using the GUI (see image below) or, using JSL, send a << Format ( ) message to the column.
Column Properties
This table lists the various field descriptors and their meanings.
JSL & Formula Editor
Custom dates are implemented by using the “Format Pattern” argument in an Informat or Format statement.
Note that “Format Pattern” is a quoted string: JMP takes care of the quotes when you’re using GUIs to build the formats, but you must include the quotes when writing custom formats with formulas or JSL.
Here is an example of JSL’s Format( ) command using a built-in (native) format. Note that the syntax of this command is Format (number, string) where string is a native format:
format(3683283977, "ddMonyyyy h:m:s");
The custom format below produces output identical to the native format above. The syntax here is Format (number, “Format Pattern”, string) where string is a collection of field descriptors and other (literal) text, like spaces, colons, etc.
format(3683283977, "Format Pattern", "
- ::");
Field Descriptors are delimited by angled brackets. Anything that isn’t enclosed in brackets is treated as plain text, allowing highly customizable output:
Tips on using Informat ( ) vs. Format ():
The Informat ( ) function is meant to convert a string into the number of seconds since January 1st, 1904. There are 3 ways that Informat () can do this:
- By having JMP guess the format. To do this, use only the original string as an argument. Because of possible ambiguities, sometimes this works, and sometimes it does not: using the code below, there is no way to tell whether the date below supposed to represent 12 March, or 3 December.
- By using a built-in format that the user specifies by adding a format string argument. Here, we specify that the date is to be represented as 12 March, not 3 December.
- Informat (“12/3/1995”, “d/m/y”)
- By using a Format Pattern:
- Informat (“2/3/1995”, “Format Pattern”, “
- //”)
Conversely, the Format ( ) function converts a number to a string, so if you use Format ( ) in a column formula, the result will be a string, not a number. To format a column of numbers and obtain a numeric result in the data table, you either must use the Column Property dialog and change the format by using the GUI (see image below) or, using JSL, send a << Format ( ) message to the column.
Examples:
Open Sample dates with conversions.jmp. This table contains many examples of Format Patterns in use.
The table below describes each column’s contents:
|
Column
|
Column Notes
|
|
Raw date text
|
The string (character data) that we wish to convert to a continuous date.
|
|
Format Pattern to be used with Informat
|
The Format Pattern that is needed to perform this conversion.
|
|
Continuous Result
|
A formula column that uses the first 2 columns as inputs, resulting in an unformatted date (the number of seconds since 1 January, 1904.)
|
|
Result displayed with a standard format
|
The date, displayed with a built-in format. (Column is continuous)
|
|
Result displayed with a Format Pattern
|
The date, displayed with a custom format. (Column is continuous)
|
|
Result converted to string using a Format Pattern
|
Here, we use a column formula and the Format() function, which converts the continuous date to a custom-formatted string. (Column is character)
|
|
Fractional Seconds
|
“Yes” if result has fractional seconds.
|
Graph Builder
The way in which the dates are displayed in the graph builder may differ from the way they are displayed in the data table, for reasons of format or necessary details. To adapt this, use the same tools as in the column properties. found in the properties of the axis in question.
The way in which the dates are displayed in the graph builder may differ from the way they are displayed in the data table, for reasons of format or necessary detail. To adapt this, use the same tools as in the column properties. You can access them in the axis properties, accessible by double-clicking on the axis in question.
In addition, there are two ways of grouping dates together: binning or wrapping.
- Binning proposes to group together all the elements of the same month, or of the same time according to the date and its scheduling. For example, if I request a binning by week: all the data for each calendar week will be grouped together.
- Wrapping allows you to group together all the elements of the same month or time according to the characteristics of the date. For example, if I request wrapping by week, the data will be grouped by day, Monday, Tuesday, Wednesday, etc. This allows me to see if there is an effect of the day on my data.
