Cumulus 3 (MX) beta documentation

This Wiki page applies to both flavours of Cumulus currently available.

Introduction

This page is about parameters used for modifying Cumulus web tags.

A Cumulus Template File is the name given by Steve Loft to any files that contain web tags. Each of those files need to be processed before they actually include values:

• MX avoids using these template files, the code itself internally generates most of the data files that are sent externally.
• The legacy Cumulus 1 software, and MX releases up to release 3.9.7 - build 3107, used lots of template files (see Customised templates.

For Cumulus MX, there is still one Cumulus template file, the web tags that supply values to the various tables in the Default Web Site are stored in websitedataT.json file. Most of those web tags use the default output format, but a few use some of the #Output modification parameters listed below. To customise the default web site, you might want to edit websitedataT.json, by using information found on this page.

To set context, let us learn the terminology with cross-references to where those features are explained further.

What is a web tag?

Put simply, a web tag is included in a Cumulus template file to indicate where Cumulus should insert values when it processes that template and produces an output file.

The output file can be:

• a web page,
• a JavaScript Object Notation (.json) file
• a JavaScript file,
• a PHP script file, or
• a eXtensible Mark-up Language (XML) file.

General Format for Web Tags

In the position in the template file where Cumulus is to insert the relevant data, place a web tag in the general format specified here:

<#tag_name [optional input selection parameters] [optional output modification parameters]>

Case sensitivity for tag names

The tag_name in the general format above is case sensitive, so please type the tag name exactly as shown in the tag name columns in the tables on the tag names page.

What is a web tag parameter?

Now we get to the terminology for what this Wiki page will document.

The parameters shown in the general format above are of two kinds:

1. Input modifying
2. Output modifying

You can include both optional input modification, and optional output modification parameters

• As the general format above shows, you separate them with spaces, e.g. <#ByMonthTempHT mon=7 format=hh:nn>.
  • In that example, the time only is returned for the highest ever temperature in July, after processing by Cumulus of the time-stamp web tag.

Case sensitivity for parameters

The optional input modification parameters always use lower case, so please type them exactly as shown in the section below.

The content of optional output parameters are only case insensitive when used in Cumulus 1.

For Cumulus 2 and later, so this includes MX, the output parameters are case sensitive and also dependent on what other output formatters are being used if any, so please read the sections on output parameters and study the examples in the tables carefully.

Input modification Parameters

Most web tags do not require any input parameters. Luckily, where they are needed, it is quite simple to use them, see table below.

• An input parameter is used where the same web tag can represent a value for a number of different past time instants.
• Each of those past time instants is represented by a different value for the input parameter.
• So a combination of web tag name and input modification parameter lets Cumulus select the value you want to see.
• The web tags that can use input modification parameters will depend on which Cumulus release you are using

Tag names	Values Available	Input Modification Parameters	Introduced	Examples	Description
Tag names listed at Table of Recent History tag names available (see Recent history page for explanation)	One value for each minute in last 7 days	d specifies number of days ago, h specifies number of hours ago, and m specifies number of minutes ago. You can use any combination of the three parameters. The same d, h, and m, parameters are used by Cumulus 1 and MX.	Cumulus 1.9.3 beta build 1033 (remain available in MX)	Examples for outside temperature: <#RecentOutsideTemp m=1> will give the temperature one minute ago, <#RecentOutsideTemp h=1> will give the temperature one hour ago (as will <#RecentOutsideTemp m=60>). <#RecentOutsideTemp d=1> will give the temperature one day ago. <#RecentOutsideTemp d=1 h=1 m=1> will give the temperature one day, one hour and one minute ago.	All values supplied for parameters must be whole numbers. If you don't supply any parameters, the result is undefined for Cumulus 1, and an illegal web tag for MX. Beware: If you use <#RecentRainToday d=2> remember that rainfall can accumulate during a day, so "d=2" returns an estimate of the rain between rollover 2 days ago and the same time as now 48 hours ago, it does not return the total rainfall 2 days ago! 'Please note: Some Cumulus users say that using <#RecentOutsideTemp d=1 m=1> is more reliable at getting the temperature at a similar time the day before, the extra minute apparently gives better results when you might not be using Cumulus all the time, or your weather station might have some drift on when it supplies readings. See which works best for you.
monthly all-time extreme records	These exist for all occurrences of the current month, and for all occurrences of each month	mon=N where N is the index of the month of the year that you want the value for (1 =January, and so on, to 12 =December)	Cumulus 1.9.3 beta build 1033 (remain available in MX)	<#ByMonthDewPointH mon=3> is highest monthly dew point for any March and <#ByMonthDewPointHT mon=3> is the related time and date. <#ByMonthTempH mon=3> gives highest temperature in any March, <#ByMonthTempHT mon=3> gives the date and time for that highest temperature	Only one input parameter applies: The value of "N" supplied should be an integer between 1 and 12 If you don't supply an input parameter (or supply an invalid value like zero) the current month will be used. Use without an input parameter applies if you want to write a template that will always supply values for the current month and don't want to process a script, to calculate the correct input parameter, before Cumulus processes the template.
Only <#SunshineHoursMonth> and Only <#SunshineHoursYear>	Values available for current month/year, and for past month/year	Listed web tags take: r=-ww (note minus sign and up to 2 digits) Monthly tags also take: m=N y=nnnn (N can be 1 to 12, nnnn is 4 digit year) Yearly tags also take: y=nnnn Omit input modification parameter to get value for current month/year	MX release 3.12.0	Monthly examples: <#SunshineHoursMonth> gives total sunshine hours since 1 minute past midnight at start of current month <#SunshineHoursMonth y=2021 m=1> for the January 2021 total <#SunshineHoursMonth r=-1> for last month's total <#SunshineHoursMonth r=-12> for same month as current month, but one year ago Yearly examples: <#SunshineHoursYear> gives total sunshine hours since 1 minute past midnight on New Year's Day <#SunshineHoursYear y=2019> for the total for 2019 <#SunshineHoursYear r=-2> total for the year before last (if current year is 2021, that returns total for 2019 as previous example)	Returns the sunshine hours total in selected period (You need a sensor to be monitoring this throughout selected period)

Output modification parameters

This page does not tell you which web tags fall into each of these 3 types:

• A few web tags always need an output format specifier
• Some web tags ignore any output format specifier as they have a fixed output format
• The majority of web tags have a default output if there is no output format modifier, but accept an output format parameter, so you can change what they output.

To make life more complicated, the availability of output format parameters for particular web tags is dependent on which Cumulus release you are running. There is a general discussion about applicability, but that needs updating as it does not specify dependencies for individual web tags.

If you are using MX:

• if your locale specifies that integer and decimal parts of real numbers are separated by a comma, there is an output parameter to replace that decimal comma by a decimal point for any script that does not recognise decimal commas
• there are two output modifiers for changing number of decimal places
• there are multiple output modifiers for changing date and/or time format

If you are using the legacy Cumulus (or a very early MX release), please skip to #Two Output (format modifier) parameters for decimal places as the changing decimal comma into decimal point parameter is not available to you.

Each of these will be explained in turn.

Output Modification Parameter for changing any decimal comma into a decimal point

General format: <#tag_name rc=y>

This only applies:

• if the web tag name represents a real number with integer and decimal parts
• if you are using MX
  • From beta release 3.0.0, build 3047 (3 February 2019), up to and including release 3.5.3, only implemented on a few new web tags (#MoonPercent, #MoonPercentAbs, #MoonAge)
  • From release 3.6.6 onwards (1 June 2020), it was extended to other web tags that output real numbers.
  • From release 3.10.5 onwards (29 March 2021), the use of <#tag_name rc=n> became also possible, to ensure decimal comma shown when locale specifies it

This output modification format parameter can be used to replace all commas in the output by a full stop (don't worry, MX does not use a comma for separating off thousands, so it is the decimal comma that becomes a decimal full stop like character when this remove comma specifier is used).

Parameter	Explanation	Example
rc=n	This is the default, so does not need to be specified. The output from the web tag will use either decimal comma or decimal point as specified by the locale in which MX is running For more information about how the computer determines whether decimal commas is your default, see #Locale section later.	Both <#tempYH> and <#tempYH rc=n> will return yesterday's highest temperature using what is specified by locale to separate integer and decimal parts
rc=y	the attribute rc takes the value 'y' to replace any commas defined by the locale with full stops to separate integer and decimal parts of the output value.	<#tempYH rc=y> will return yesterday's highest temperature as integer part then full stop then decimal part, regardless of local

Why would you want to remove decimal commas? Well because the JavaScript language cannot understand decimal commas, and MX has several scripts written in this language, equally some third party alternative web pages rely on ajax to update them (and Ajax uses JavaScript).

Controlling the number of decimal places

Internally, Cumulus stores numbers in binary. You cannot represent base 10 decimal places exactly in base 2.

Therefore, Cumulus stores to a precision that would generally give about 24 significant figures when expressed in base 10.

Cumulus is written to assign particular numbers of decimal places to any outputs it makes, and in any logging of current, or extreme values (for day or longer periods). It determines these precisions, by reference to the units chosen for outputs.

From release 3.12.0, you can set the default number of decimal places to output for all derivatives of temperature, pressure, etc. by advanced settings. Those settings can force output as integers, stopping these modification parameters from working.

As handling of each web tag is coded individually, the number of decimal places output by default in any web tag might vary slightly from the above default. As Cumulus has been developed, people have commented that these defaults do not reflect the precision of their instrumentation (weather stations used with Cumulus tend not to have the accuracy of those used by meteorologists, or are not recalibrated as often).

This section (and its subsections) only applies to tag names that output real numbers (with integer and decimal parts). You can't change anything that is output as an integer, or is text with these parameters, nor can you change the decimal places for any time element.

Consequently, gradually Cumulus has allowed more and more of its output to take an output format modifier that allows people to control number of decimal places shown.

Two Output (format modifier) parameters for decimal places

From release 3.10.5 (which did a big rewrite of web tag handling), you can modify real number output for individual tag names, using output modification parameters in either of the following formats:

• <#tag_name dp=i> and
• <#tag_name tc=y>

These can be applied to any tag names that represent real numbers (with integer and decimal parts).

Rounding to a specific number of decimal places

dp=i is used for both Cumulus 1 and MX.

• The value i following the attribute dp is an integer, it represents how many decimal places you want for the output you see.

This functionality was trialled in the original Cumulus, but has been properly implemented in MX.

If you are using the legacy Cumulus (1.9.4), only <#latitude dp=i> and <#longitude dp=i> were able to be output with "i" denoting number of decimal places, e.g. <#latitude dp=5> gives "59.24250".

If you are not using latest MX release, you may find this is not available for particular web tag names

1. From beta releases (3.0.0) onwards, <#latitude dp=i> and <#longitude dp=i> were able to be output with "i" decimal places
   • But this output modification parameter could not be applied to any other tags in the MX beta.
2. MX when it came out of beta, added this output modification parameter usage in the moon tags <#MoonPercent> and <#MoonPercentAbs>).
   • Specifically, <#MoonAge> gives "11" but <#MoonAge dp=3> gives "11.234"

Truncation of unwanted decimal places

This output format modifier is only available in MX.

tc=y is the truncation parameter, the attribute tc takes the value 'y' to remove decimal places by truncation. e.g. <#MoonAge tc=y>.

Whilst many people want Cumulus to round output as done by the previous parameters, there are circumstances when rounding down (or truncation) gives the result desired.

• If you are using an early release of MX, you will need to research whether this is available for particular tag names
• Later releases of MX implement this for any tag that by default outputs decimal places.

Note, truncation by MX converts real numbers to integers. There is no option to truncate to one decimal place, Airports are expected to report air field level (QFE), and sea level (QNH), pressures truncated to one decimal place rather than rounded.

Multiple Output Format Modifier parameters for times and dates

These are highly complicated, and so have been left until after the simpler ones!

To start with a simple example, suppose you want date/time in ISO 8601 format:

1. This means, something like 2019-02-28 06:59:05.
2. Take the tag name (from tables on Webtags page)
3. Next check in #Which tag names take date/time output formatting modifiers to see if that tag accepts both time and date modifiers
4. If our tag name does accept both date and time modifiers, simply modify the web tag as shown here <#tag_name format="yyyy-MM-dd HH:mm:ss"> where tag_name is set from step 1, but all the rest is typed as shown.
5. To explain each element in that format value, look in #Year formats, #Month formats, #Day formats, #Use of spaces, #Time formats.

Should you want a different date/time format, then the sub-sections just referenced should help you to select a different arrangement, although there are some more options in #Date formats.

Which tag names take date/time output formatting modifiers

There are nearly a thousand different tag names.

There are a few time-related tag names labelled as being fixed format, that means any time/date output modifiers you might try would be ignored.

Some tag names, although it can be hard to tell which, report durations. You can use time modifiers to change how these are reported. Examples include <#daylength format=H:mm:ss> and <#MonthDailyRainHD format=H:mm>.

Time modifiers can also be used to change the way that clock times are reported, you might only want the hour, or you might (if resolution is available) want to also see seconds. The webtags page has columns headed "Time" to clearly identify all tag names that report clock times.

There are some tag names (e.g. moon rise) that relate to an event that does not happen each Earth day, so those tags have to be able to report "--:--", and you cannot modify their output.

There are some tags (e.g. highest temperature range in month/year), for which Cumulus has been coded to report "--" for value, and "--:--" for time, on the first day of that period (because there is only a partial day to consider you might not yet have experienced a true maximum and a true minimum), so modifying the time output can only be done on subsequent days.

It is not practical to indicate which time/date modifiers are accepted on a tag by tag basis. It would involve a lot of repetition. Instead, the following table, explains much more simply, which web tags will accept time and/or date output modifiers:

Cross-reference to table on web tag names page	Tag names that accept only time output modifiers	Tag names that accept only date output modifiers	Tag names that accept both time and date output modifiers
Any tag names that don't report times nor dates	None	None	None
Webtags#No_Commas	None	None	None
Webtags#Date_.26_Time, Webtags#Day/Night/Sun/Moon	Only <#timehhmmss>, <#minute>, <#hour>, <#sunrise>, <#sunset>, <#dawn>, <#dusk>	Only <#LatestErrorDate>, <#date> (but no others)	Only <#LastDataReadT>, <#time>, <#metdate>, <#metdateyesterday>, <#update>, <#LastDataReadT>
Webtags#Today, Webtags#Yesterday	Any tag name in "Time" column of linked table	None	None
Webtags#Monthly, Webtags#Yearly	Any tag name in "Time" column of linked table in first column	Any tag name in "Date" column of linked table in first column	None of the tag names. For explanation see the ^ below this table
Webtags#All_Time, Webtags#Monthly_All_Time_Records	None (all tag names combine both time and date)	None (all tag names combine both time and date)	Any tag name in "Date/Time" column of linked table
Webtags#Davis	None	<#StormRainStart> only	None

^ For the monthly and yearly web tags, the date and time are in separate tag names. It is not possible to get both time and date out of either tag name.

Note: There are some monthly/yearly web tags (e.g. wettest day) where a date tag is available (i.e. <#MonthDailyRainHD>), but there is no time tag. As explained before, in that wettest day example <#MonthDailyRainHD format=H:mm> returns the duration in hours and minutes since rollover for which rain continued to increase on that date, not the clock time. For rainfall, only <#LastRainTip> can have output modifiers added to report a clock time.

Locales

The default format for many tag names reporting date and/or time is dependent on the locale you are using for running Cumulus (1 or MX).

The effect of some output format modifiers is also dependent on locale.

For MX running on most operating systems (and therefore using Mono), type locale to see the default locale that will be adopted when mono-complete is installed as MX will, by default, take locale setting from Mono. When you start MX, you can ask it to use a different locale to that picked up by Mono, by adding the parameter "-lang locale-code", see examples at MX_on_Linux#Parameter_for_changing_Locale. For example, the Australian English language with UTF-8 encoding locale is defined as: en_AU.UTF-8.

The available locales on your computer in Linux are listed by locale -a. For example, the Russian locale would be selected as the one your computer uses for the current session only by using LANG=ru_RU.utf8 either typed into a terminal session before you start MX, or used as a parameter (preceded by "-") as you start MX interactively.

For permanently changing the locale used by your system, the instructions vary considerably according to the kernel used in your operating system, so you need to look up the instructions for yourself. However, if you have a graphical user interface, such as the full Raspberry Pi Operating System provides, you might have a configuration command in terminal mode and a configuration app accessed (within Preferences) from the "Raspberry" key on the official keyboard. For the Raspberry Pi, please read Raspberry Pi computer page for more details.

For Microsoft Windows Operating Systems, a Language is defined within the "Region" page of the Settings app. That should be sufficient for the legacy software that uses Delphi.

However for MX date and time formats within Windows Operating System, you must use the older Control Panel (go to "Clock and Region" screen, choose "Change date, time or number formats", choose "Language preferences") because it is only there that you can adjust all the defaults used by .NET.

Time zones

Before I explain what time/date output modifiers can do, something they can't do.

All web tag outputs are in local time, except <#timeUTC>.

Although Cumulus 2 internally stored all date/times in UTC, no flavour of Cumulus is currently able to output the time-stamp for any weather extreme in UTC, if your current time is not in UTC.

However, for MX only, you can use a script to convert a time to UTC. This is not the place to tell you how to write the script, but taking time of highest pressure today as an example, you would use <#TpressTH format=Hh:mm> and <#TpressTH format=zzz>, the first gives hours and minutes in your local time, and the second gives the offset that needs to be applied to that time to convert it to UTC.

Time resolution

For the legacy software, there may be no point in asking for seconds, as Cumulus 1 did some actions at one minute intervals.

If Cumulus obtained archive data, as part of the catch-up process it can do when it restarts, any time-stamps for that period can only be the time of a particular archive record, so that might be every half an hour, but not aligned precisely with hour changes.

Dependency on Cumulus flavour

You cannot,in general, use the same date/time formatting for both the original (legacy) Cumulus and MX.

All the tables explaining what is available, attempt to show what is used in each flavour for each type of output, both by including separate columns for each flavour, and by giving examples in each flavour.

For the legacy software

I deal with this first, just because it is simple!

From version 1.9.1, most web-tags that report any form of time or date will accept an optional 'output format' parameter (we have already seen whether this can only affect time, only affect date, or both).

The legacy Cumulus uses Delphi to interpret the output modifiers:

• For most modifiers, a particular character produces the same output regardless whether the output modification specifier is in capitals or lower-case
  • There is an exception, the case you use for any am/pm output format modifiers determines the case that is output.
• In general, the context of a modifier does not affect the output it produces
  • Again, there is an exception, "m" or "M" has two different meanings (minutes or month) depending on context.

The complications with MX

Cumulus MX works internally with dates specified in either a day before month before year format, or ISO 8601 date format where year comes first (yyyy-MM-dd) depending on context. Compatibility with the legacy software has so far meant while the *.ini Files have adopted the year first approach, the *.txt Files have stuck to date formats as used in the legacy definitions.

For Cumulus MX output formatting, the date and time modifiers are complicated by the fact that the same character can have 4 different meanings depending on its case (capital letter or lower-case letter), and depending on whether it is on its own (standard format) or with another modifer (custom format). Sounds confusing? Well it is complicated.

Consider context first:

• <#tag_name format=x>
  • If the x in the above general syntax is a single character, it represents a standard format code
  • The standard characters for dates and times are defined at standard-date-and-time-format-strings
• <#tag_name format=xyz>
  • If the xyz in the above general syntax is replaced by two or more characters, it becomes a custom format code (combinations of characters, or single characters prefixed by %)
  • The custom characters for dates and times are defined at custom-date-and-time-format-strings

Consider case next:

• Cumulus MX (when running on Windows) uses the .NET software which is provided as standard by Microsoft Windows.
  • ".NET" was originally operating system independent, later only Microsoft Windows specific components were included, but since November 2020 ".Net" is used for an operating system independent version that originally Microsoft issued under another name!
  • (actually it is possible to install and run "Mono" in Windows Operating Systems).
• If Cumulus MX is running on any Linux distribution (including Raspberry Pi Operating Systems) or Mac OS X, or any other device that uses an UNIX derived operating system, then MX uses Mono software for same purposes. (MONO is a operating system independent version of .NET, although they are developed independently, they have common origins).

Use of spaces

There are multiple symbols for specifying dates and times, and you might want spaces to appear between symbols in that output format.

You need to add quotation marks to the output format specifier if spaces are present.

The first complication is that the parser that interprets time/date characters has two ways of interpreting a space character, depending on what immediately follows. In the tables, below, I have used a "%" in various places. In any of those places, a space is not a gap between characters, but an alternative to "%". I discuss this later in #Migrating from legacy Cumulus 1 to MX section.

For a space character to be interpreted as a gap between symbols, the symbol that follows the space must include at least two characters. The syntax <#tag_name format="x y z"> works if the y and z in it are representing multi-character symbols. To explain this, an example is <#TpressTH format="h:mm tt"> as both h:mm and tt are multi-character symbols, we have inserted a space after the minutes.

If we have single character specifiers, then a space has to be enclosed in single quotes to mark it as a literal, not a modification of the subsequent single character. To explain this, a simple example is <#MonthPressHD format=" d' 'M">, here the month number is a single character "M", so to insert the space we have to treat it as a literal by enclosing the space character in single quotes and the whole specifier in double quotes.

Literals are discussed fully in the #Including literals in format parameters sub-section later. If we want to include other characters not to be interpreted by the date time parameter parser, and spaces, then both double and single quotes must be used, and the spaces must be within the single quotes. An example, that shows all the options that MX allows, with literals is <#TpressH format="\a't 'h:mm' ' tt' <small>on 'd/M/yyyy' </small>'"> .

Finally, the use of literals can cause you a problem if you want to use a date/time specification in a script because the script wants literal delimiters outside any web tags, so that delimiters remain when the web tag itself has been processed into a string by Cumulus. This means the type of quotes (single or double) used outside the web tag, cannot be used within the web tag. The complicated sounding (but actually simple solution) is to avoid placing literals, and/or spaces, within any output format specifier, instead put single quotes round the whole content. What you thought of putting as literals within any web tag is instead typed outside with separate web tags for the part of the specification before and after each literal. An example to make this clearer is $MXDateTime = '<#date format=yyyy-MM-dd>' . 'T' . '<#time format=hh:mm:ss>';, which is written in PHP Hypertext preprocessor format, the literal 'T' has been inserted between two separate web tags.

Year formats

These are the simplest output format modifiers. We choose from 2 options, and because both involve more than one character their context does not matter. Although the legacy Cumulus will accept upper case as meaning same as lower case, it is simplest if we just show the lower case options needed for MX:

Specifier	Displays	Example
yy	Displays the year as a two-digit number (00-99).	19 produced by <#LastDataReadT format=yy>
yyyy	Displays the year as a four-digit number (2000-9999).	2009 produced by <#LastDataReadT format=yyyy>

Month formats

All locales offer both numerical and alphabetical formats for representing months.

• In the following table "MMM" is shown as producing short month name.
  • What language is used, and what characters appear, depend on what is set up for your language in your settings (by default or by you changing your settings)
  • In British English (UK) locale this will be the appropriate 3 letter abbreviation that starts with "Jan" and runs to "Dec"
  • It appears that language settings in many locales (not "en-gb"), add a full stop to any abbreviations

and in that case the 3-letter "MMM" is turned into a 4-letter output (e.g. Australia settings default would output from "Jan." to "Dec.")

• • • MX has been coded to remove that full stop in various places (like in standard log file naming and NOAA report naming), but at the time this section was edited, "MMM" still reports the full stop in a web tag output if your locale uses it
• In the following table "MMMM" is shown as producing the full name for a month
  • The output you get will depend on the language defined in your locale
  • In English locales, the output will be in the range "January" - "December"

Delphi Specifier for Cumulus 1.9.x	Mono/.NET Specifier for Cumulus MX	Displays	Example
M (or m)	%M	Displays the month as a number without a leading zero (1-12). Cumulus 1.x.y:If the 'M' or 'm' specifier immediately follows an h, hh, HH, or H specifier, the minute rather than the month is displayed. Cumulus MX: Note that including a ' ' (space) or '%' before the M makes it a custom modifier, so different from the format=M included in previous table.	2
MM (or mm)	MM	Displays the month as a number with a leading zero (01-12). Cumulus 1.x.y:If the 'm' or 'M' specifier immediately follows an h, H, HH, or hh specifier, the minute rather than the month is displayed.	'03' produced by <#LastDataReadT format=MM> or <#metdate format="MM">
MMM (or mmm)	MMM	Displays the month using the strings defined for short month name in the Locale.	'Jun' produced by <#metdate format="MMM"> (English locale)
MMMM (or mmmm)	MMMM	Displays the month as a full name using the strings appropriate to the Locale.	'June' produced by <#metdate format="MMMM"> (English locale)

Day formats

All locales offer both numerical and alphabetical formats for representing a day.

The table below relates just to the day part of any date specifications.

}Delphi Specifier for Cumulus 1.9.x	Mono/.NET Specifier for Cumulus MX	Displays	Example
d (single character)	%d	Displays the day as a number without a leading zero (1-31). Note that Cumulus MX requires a ' ' (space), '%' or another modifier to be included, as 'd' on its own is inconsistent - see #Date formats table).	27 produced by: <#metdate format="d"> <#metdate format="%d">
dd (2 characters long)	dd	Displays the day as a number with a leading zero (01-31).	07 produced by <#metdate format="dd">
ddd (3 characters long)	ddd	Displays the day as an abbreviation (in UK English will output Sun-Sat) using the strings appropriate to the Locale. As for month above, the short day names that are generated depend on your locale, so you might see additional punctuation defined for abbreviated names in some locales.	'Wed' produced by <#metdate format="ddd"> (English locale)
dddd (4 characters long)	dddd	Displays the day as a full name (Sunday-Saturday) using the strings appropriate to the Locale.	'Friday' produced by <#metdate format="dddd"> (English locale)

Date formats

The #Year formats, #Month formats, and #Day formats listed above can be combined to make up a date output modifier, but there are some other modifiers available that can produce whole dates.

#Locales will define a Short Date Format and a Long Date Format. You will see references to those in the table below explaining available output format modifiers, for example the single character output format modifier (G or c) listed at the start of the table below.

If you are in the USA, Cumulus will not use your month first date internally or in any files in the data sub-folder, but you can see your preferred format asan output from web tags, as you can can combine the month specifier, with the day specifier, in that order, to get an output where the month appears first (see example in table below).

Please could an American contributor please check if the "M" modifier works for their locale as shown and update the table below.

Delphi Specifier for Cumulus 1.9.x	Mono/.NET Specifier for Cumulus MX	Displays	Example
c	G (as single character format)	Displays the date using the format given by the Short Date format, followed by the time using the format given by the Long Time format. The time is not displayed in Cumulus 1 if the date-time value indicates midnight precisely.	'22/03/2019 09:47:25' produced by <#time format=c><#time format=G>
ddddd (5 characters long)	d (as single character format)	Flavour issues: Cumulus 1.x.y: 'ddddd' will output the date using the format given by the Short Date format. This MX parameter (when on its own, without space or '%' prefix) displays inconsistent behaviour as its effect depends on the tag name (sorry, there is no definitive list from the MX author specifying effect by individual tag name) with which it is used (you will need to experiment for yourself and compare with the two examples).	'22' returned by e.g. <#yesterday=d> Short date format e.g. '22/03/2019' (British Locale) produced by: <#metdate format=ddddd> <#metdateyesterday format=d>
dddddd (6 characters long)	D (as single character format)	Displays the date using the format given by the Long Date format. The MX 'D' parameter cannot be combined with any other parameters.	e.g. '22 March 2020' (British Locale)
"D MMMM YYYY"	D	Long date format	e.g.4 December 2009
"D MMMM"	M	Day of month followed by Month name (except USA locales). Compared with next table where format=%M used, format=M on its own returns both Month and Day according to local format	e.g. 22 July (English Locale)
"MMMM D"	"MMMM d" if M alone does not work	USA format of month before day of month	e.g. July 4 (USA format)
TT	T (as single character format)	Displays the time using the Long Time format. Note that this is a full time specifier and "T" is on its own as we are using a single character format.	'09:47:56' (might not use colon in your locale) produced by <#LastDataReadT format=TT> <#LastDataReadT format=T>

Time formats

Here context matters, so both standard (single character) and custom (two or more characters) formats are shown in the following table. As explained earlier, time formats can be used with durations and clock times.

In some rows of this table, square brackets [] indicate optional items, they are included just to make it clearer how items can be combined in a single output parameter. If you want to include what is shown in square brackets you don't type the square brackets e.g. <#LastDataReadT format="h:nn am/pm">

Delphi Specifier for Cumulus 1.9.x	Mono/.NET Specifier for Cumulus MX	Displays	Example
h	%h	Displays the hour (12 hour clock) without a leading zero (1-12)	7
h AM/PM	h tt	Displays the hour (12 hour clock) without a leading zero (1-12) in combination with AM/PM. For Cumulus 1 the formats for am/pm depend on the case in which you type the parameter as shown later in this table What "tt" produces depends on locale settings for your device, it might be capitals or it might be lower case (in Windows use Control Panel, not Settings app, to get to these regional additional settings).	7 PM
h:nn [AM/PM]	h:mm [tt]	Displays the hour (using 12 hour clock) without a leading zero (1-12) followed by 2 digit minutes [optionally in combination with AM/PM whose case varies as explained in previous entry]. For Cumulus 1, the minutes can be represented by 'mm' (instead of "nn") only when appearing in combination with 'h'	'10:27 am' produced by <#LastDataReadT format="h:nn am/pm"> <#LastDataReadT format="h:mm tt">
H (or H)	%H	Displays the hour part of any duration or clock time, without a leading zero (0-23). Note that '%' before the "H", this makes it a custom modifier, needed because H is on its own.	7 produced by <#daylength format=H> <#daylength format=%H>
H:mm (or H:nn)	H:mm	Displays the hour part of any duration or clock time, without a leading zero (0-23) followed by 2 digit minutes for that duration or clock time. Note that %, of previous example, is not needed as the H is not on its own.	'6:27' or '17:49' produced by <#LastDataReadT format="H:mm">
HH (or hh)	HH	Displays the hour part of any duration or clock time, using 24 hour clock with a leading zero (00-23).	'06' or '17' produced by <#LastDataReadT format=HH>
hh [am/pm]	hh [tt]	Displays the hour part of any duration or clock time, (12 hour clock) with a leading zero (01-12) [optionally, if it is a clock time, in combination with am/pm]. For Cumulus 1 the case output for the optional 'am/pm' depends on the case used for that parameter as shown later in this table For MX, the optional 'tt' displays the contents of the device locale setting for AM string for midnight until any hour before noon, and the contents of the PM string for noon or any hour after noon but before midnight	'07 am' produced by <#LastDataReadT format="hh am/pm"> <#LastDataReadT format="hh tt"> if locale specifies lower case
hh:mm (or hh:nn or 'HH:NN') [am/pm]	hh:mm [tt]	Displays the hour part of any duration or clock time, (12 hour clock) with a leading zero (01-12) followed by 2 digit minutes [optionally, if it is a clock time, in combination with am/pm]. For Cumulus 1, the minutes can be represented by 'mm' only when in combination with 'h', in other contexts 'mm' is interpreted as month number, and the case output for am/pm depends on the case used for that parameter as shown later in this table. As Cumulus 1 is case insensitive there are variants with capital letters available. For MX, the optional 'tt' displays the contents of the device locale setting for AM string for midnight until any hour before noon, and the contents of the PM string for noon or any hour after noon before midnight	'8:27 am' produced by <#LastDataReadT format="h:nn am/pm"> <#LastDataReadT format="h:mm tt">
n	%m	Displays the minutes of any duration or clock time, without a leading zero (0-59). As other examples show, the % is only needed when "m" is on its own.	7 produced as a duration in minutes by <#daylength format=n> <#daylength format=m>
nn	mm	Displays the minutes of any duration or clock time, with a leading zero (00-59).	'07' produced as a duration in minutes by <#daylength format=nn> <#daylength format=mm>
s	%s	Displays the seconds for any duration or clock time, that has resolution to less than a minute, without a leading zero (0-59). As other examples show, the % is recommended when "s" is on its own, although I have not found any alternative meaning for "s" on its own.	9 produced by <#metdate format=s>
ss	ss	Displays the second with a leading zero (00-59).	'06' or 19 produced by <#LastDataReadT format=ss>
z	FFF	Displays the millisecond without a leading zero (Cumulus 1: displays 0-999, Cumulus MX: displays either nothing, or displays 1-999, so don't write any code that assumes the MX output is numeric). Note that the system clock (before Windows 10 64-bit systems) only has precision to 15 ms, so don't use this modifier if your Cumulus is running on an old version of Windows.
(not available)	ff (or f)	Displays hundredths of a second (or tenths) with leading zero(s)
zzz	fff	Displays the millisecond with a leading zero (000-999). Note that the system clock (before Windows 10 64-bit systems) only has precision to 15 ms, so don't use this modifier if your Cumulus is running on an old version of Windows. The 'fff' modifier in MX can actually be extended to 'ffffff' for output to a millionth of a second!	09:47:25.000' produced by <#time format=hh:nn:ss.zzz> <#time format=hh:mm:ss.fff>
(not available)	zzz	Displays the offset of any time from UTC in hours and minutes	e.g.-07:00
(not available)	"h:mm K"	Effectively another way of including time zone after a time, but it can only be used for times not in UTC (if I understand correctly)	(no examples supplied yet)
t	%t	Displays the time using the Short Time format. Remember that 't' combined with other specifiers (or preceded by space or '%') has a different meaning - see below.	'09:47' produced by <#LastDataReadT format=t> (might not use colon in your locale) for both flavours of Cumulus
am/pm or Am/Pm or AM/PM	tt	Uses the 12-hour clock for the preceding h or H specifier, and displays 'am' for any (clock time) hour from midnight until just before noon, and 'pm' for any hour from noon onwards. The am/pm specifier for Cumulus 1 can use lower, upper, or mixed case, and the result is displayed accordingly. For MX, 'tt' displays the contents of the device locale setting for AM string for midnight until any hour before noon, and the contents of the PM string for noon or any hour after noon before midnight, so whether it displays in capitals or lower case is determined by the locale settings, not the case of "tt".	{Version badge 1}} 'am' produced by <#LastDataReadT format=am/pm>, 'AM' produced by <#LastDataReadT format=AM/PM>
h a/p	h t	Uses the 12-hour clock for the preceding h or H (clock time) specifier, and displays 'a' for any hour from midnight until before noon, and 'p' for noon or any hour after noon. The a/p specifier can use lower, upper, or mixed case, and the result is displayed accordingly. whether it displays the "a" or "p" in capitals or lower case is determined by the locale settings, not the case of "t".	see previous example
ampm	(see above for 12 hour formats)	This displays the contents of the device locale setting for AM string for midnight until any hour before noon, and the contents of the PM string for noon or any hour after noon before midnight. {{Version badge 1}Uses the 12-hour clock for the preceding h or H specifier	see previous examples
/	/	Displays the date separator character given by the Date Separator. It might not display a slash.	'/' for typical British locale
:	:	Displays the time separator character given by the Time Separator. With Cumulus 1, this might not display a colon. Note that by default Cumulus MX expects a locale to use ":" for any time separator.	':' for British locale

Some Extra Information

Having covered the basics of both date and time modifiers above, it is time to talk about incorporating other information in an output modifying date/time format specification.

Basically, we can include literal characters, and we can include HyperText Manipulation Language tags, in our specifiers.

Finally, there will be a section on migrating from the legacy Cumulus to MX and how to modify the web tags in your templates to keep them working.

Including literals in format parameters

#Use of spaces explained how double quotes were needed for date/time output specifiers containing spaces. It briefly talked about including literals, and we will expand on that now.

Consequently, you cannot include double quote characters in any other position (see here for work-around).

You should put anything that is additional, to the defined format modifier specification below, into single quotation marks to prevent it being interpreted as a date or time format modifier. In MX, such single quotation marks should include the spaces round the additional literal text.

1. For example, the word "on" contains the character "n", which for Cumulus versions 1.9.1 to 1.9.4 will be interpreted as a time format modifier unless you put it into single quotation marks. Example of valid Cumulus 1 syntax: <#TtempH format="'at' hh: mm 'on' dd / mm / yyyy">.
2. You can include HTML tags (but they cannot have any attributes because both single and double quote characters have defined meanings) and special characters as quoted text within the 'format' parameter.
   Example of valid syntax: <#TapptempH format="'at 'h:nn' 'am/pm '<small>on' d/m/yyyy'</small>'">.
   • See next sub-section for more information on incorporating HTML if you are using MX.

Note for MX - you can use single quotation marks round spaces and text (e.g. ' on '), but you can also use '\' as escape character (e.g. for 'on' use \o\n). However for at the only alternative is \a't' because the character t has another meaning and escape followed by a "t" i.e. "\t" becomes a tab!

Delphi Specifier for Cumulus 1.9.x	Mono/.NET Specifier for Cumulus MX	Displays	Example
'xy'	'xy' or \x\y	Characters enclosed in single quotation marks are displayed as such, and do not affect formatting. In MX each character to be displayed as it was typed can be prefixed by a backslash. Also remember that any spaces in a MX modifier might need to be within single quotes as space is also used to change what a modifier represents. I told you MX modifiers were more complicated!	Hyphens are added in this PHP language example '<#LastDataReadT format=yyyy>'.'-'.'<#LastDataReadT format=MM>'.'-'.'<#LastDataReadT format="dd">'

Using HTML tags within format parameters (available in MX only)

Example using a class to change the look of part of the output

<#TapptempH format="dd' 'MMM' 'yyyy'<span class=\'xx\'> at 'HH:mm'</span>'">

the output from this will look like 04 Dec 2018 at 10:12

Note where the quotes are, and where you need to use '\' escape characters.

Example using HTML tags

<#RecentTS d=2 format="h:mm' 'tt'<small>on' d/M/yyyy'</small>'">

This puts the date in a smaller font than the time

Migrating from legacy Cumulus 1 to MX

If you have created any legacy cumulus template files, then in each template, you will need to do some editing. Everywhere a web tag appears with an output modifiers that is used to specify a date and/or time format, has to be edited before that template will work for MX.

Here are the main reasons:

• the reserved characters are different in C1 and MX (affecting use of literals like "on" and "at" that appear in many English time-stamps)
• the Delphi in legacy Cumulus is case insensitive, so for example "H" and "h" have the same meaning
• MX is case sensitive, and symbols mostly have different meanings when one symbol is used to when that symbol is used with others, so for example "H" and "h" have different meanings, and if not used with other symbols will need to be preceded with a "%" to have same meaning as they have in combination with other symbols
• In the legacy cumulus, a symbol like "d" has the same meaning for any tag
• MX is inconsistent, a symbol like "d" changes its meaning depending on the tag it is used with (e.g. the script conditional '<#metdateyesterday format=d>' == '<#yesterday format=d)>' will never be equal as the LHS returns a full date and the right hand side returns day of month only)
• the symbols used for representing such modifiers as minutes, month, am/pm, are different between C1 and MX.
• MX introduces the concept of escaping characters (a \ placed before a character can be either a control sequence or an instruction to display the character)
• In the legacy Cumulus, a space is a gap between characters
• In MX, a space must be within a literal, as a space before a symbol has the same effect as "%", (it changes the interpretation of a modifier character).

Confused even more now? I'm not surprised, but maybe some examples will help.

Examples

• Examples related to case selection
  1. In Delphi, "nn" means "minutes" for Cumulus 1, but "minutes" is "mm" for .NET or MONO in Cumulus MX.
  2. The hour in 24-hour format with leading zero, in non case sensitive Delphi (Cumulus 1) 'HH' or 'hh' would be treated as same, but in .NET or MONO it must be "HH" (Cumulus MX).
  3. The hour in 24-hour format without leading zero, in non case sensitive Delphi (Cumulus 1) 'H' or 'h' would be treated as same, but in .NET or MONO it must be "%H" (Cumulus MX).
  4. For 12-hour specifiers, please see the table, as this is far more complicated.
• You might be put off by references within .NET and MONO (Cumulus MX) to single/standard characters and custom modifiers, the following 3 examples may add clarity:
  1. For example, <#MonthTempHD format="d"> is a single character format modifier, therefore the 'd' acts as a standard modifier, and causes for a date of 22 July 2014 for the highest temperature in the month to be returned in the standard short date format e.g. '22/07/2014' (exact contents for any one date vary by locale).
  2. Similarly, <#MonthTempHD format="M"> is a single character format modifier and therefore the 'M' acts as a standard modifier and causes the date for the highest temperature in the month to be returned in the standard day and month format e.g. '22 July' (exact contents for any one date vary by locale).
  3. Whilst <#metdate format="d M"> is not a single character format modifier and therefore both the 'd' and the 'M' are interpreted as custom modifiers and cause the current date to be returned as a digit(s) for the day and a digit(s) month (in a without leading zeroes format) e.g. '6 7' would be returned for 6 July.
  4. Alternatively, <#MonthTempHD format="%d"> is NOT a single character format modifier, therefore the 'd' acts as a custom modifier, and causes a date of 22 July 2014 for the highest temperature in the month to be returned as the day of the month only '22' in all locales.
  5. Similarly, <#MonthTempHD format="%M"> is NOT a single character format modifier and therefore the 'M' acts as a custom modifier and causes the same date for the highest temperature in the month to be returned as the month number '7'.

In both Cumulus 1 and MX if you want a space character within your output, the output specifiers must be enclosed in double quotes. If that space character is next to a non modifier (e.g. around word "at") then the single quote needing to surround the at should be widened to include the spaces in MX, but Cumulus 1 does not care if single quotes excluded spaces. However, with MX, single quotes enclose multiple characters, but there is an alternative way to deal with some single verbatim characters to cover next.

So let us compare these two alternative ways that MONO and .NET escape any characters that are not being used as format specifiers.

• In Delphi you can put the 'verbatim' characters inside single quotes (Cumulus 1); this is often used to (in English) include words like ' on ' and ' at ' in the formatted output.
• in .NET or MONO you can still use single quotes (as mentioned above extended to include adjacent spaces),
  • but alternatively you can escape each verbatim character with a backslash as prefix (Cumulus MX).
• You may need to use both single quotes and back slashes in some format specifiers, depending whether the characters you want to include can be interpreted as control characters (yes, backslash is also used to escape control characters, so backslash will NOT work for some characters such as those in "on" and "at" [\n will produce new line not the letter n, \t will produce a tab not the letter t]), consequently for some characters you must use the literal approach to include them in your format.

Past history for this page

This page is a complete redesign of how to present information that was previously on the Webtags page, so look there for past content by selecting "history" tab.

Trying to make the old design made for the original Cumulus software, work for MX which is now very different, made the old page unwieldy.

Forum reference

Steve Loft published a table showing comparison between output date modifiers for Cumulus 1 and MX at Cumulus MX forum. The table there was based on the table that appeared in this Wiki when only the original Cumulus existed, so it was designed to help people migrate to his MX beta, it was not intended as a definitive list of what modifiers were available for MX (Steve instructed people to look them up on some Microsoft sites).

The subsequent comments in the forum suggested his layout got people confused. Most of that confusion came in two circumstances:

• When someone wanted to use one date or time modifier on its own
• When someone who had been using Cumulus 1 swapped to MX and wanted to replace a combination of output modifier characters that was not explicitly shown in his table.

That all comes from the fact that when a MX modifier consists of a single character it can mean something different to when it appears with other characters.

In Cumulus 1, "m" or "M" had two meanings depending whether it was combined with "H" or "h" (when it represented minutes), or on its own or with any other code (when it represented month). But for Cumulus 1, there is no other case where it matters what context a modifier is put in by the use of other modifiers, and no other modifier takes more than one meaning.

In MX it is much more complicated, to take a few examples "D", "H", "M" represent different items on their own to what they represent when combined with other characters. That other character can be as simple as using a space or a "%" to modify the meaning of the character.

Looking at the tables, now included above, you can see "G" is used on its own because it represents a full date-time specifier. "D" is similarly used on its own represents the long date format. If we only want the day of month number we must use "%d" to avoid the meaning of short date format that "d" on its own represents.

If we want the typical Cumulus date-stamp of day of month number and month, then we have two choices, because both "d M" and "M" will work. This illustrates how "M" has a different meaning on its own and with another modifier.

Hopefully, the way that information is now presented on this page makes any use of parameters for web tags much easier now.