data filtering lib

For 18 year old women:

age = 18
gender = "female"

For men, 18 - 35:

gender = "male"
age >= 18
age <= 35

(this is what we'll use for MP)

You can do OR rules instead of AND rules using "[]"

People who are 18 years old or 21 years old:

[
  age = 18
  age = 21
]

Women who are 18 or 21:

gender = "female"
[
  age = 18
  age = 21
]

This also works:

gender = "female"
[ age = 18, age = 21 ]

"{}" means "AND"

Women who are 18 or 21, or between 35 and 55:

gender = "female"
[
  age = 18
  age = 21
  {
    age >= 35
    age <= 55
  }
]

Quotes around the Field names are optional when the name is only letters, numbers, dashes, and underscores. The following three examples are all exactly equivalent:

gender = "female"

"gender" = "female"

'gender' = "female"

Maybe applied to either "AND" or "OR". Used with "OR" will match "all those not being 18 or 21"

![
  age = 18
  age = 21
]

When used with "AND" it will match all except women being 18

!{
  age = 18
  gender = "female"
}

Allows "IN" lookup. So:

age in (20, 30, 40)

... is the same as:

[
  age = 20
  age = 30
  age = 40
]

"NOT IN" may be applied in the following way:

age !in (20, 21, 22)

and it will match all the people except those being 20, 21 or 22 years.

Different formats are allowed, so the following 2 examples are equivalent as the above, either separated with a newline:

age !in (
  20
  21
  22
)

or mixed:

age !in (
  20, 21
  22
)

Apart from integers, array may contain decimals or string as long as all its elements are of the same type:

"price" in (9.95, 10.95, 11.95)

or

"color" in ("blue, "red", "green)

This WOULDN'T work:

age in (20, "21", 22, 23)

An empty "AND" block will match all users

{}

An empty "OR" block will match no users

[]

This means you can set {} as your last filter in cases where you want to match various groups and then have an "everybody else" group.

Numbers:

• "x" = 1
• "x" = 500000
• "x" = 7.5
• "x" = -1642
• "x" = -0.9141

Strings:

• "x" = "hello"
• "x" = "a complete sentence"
• "x" = 'this time with single quotes'
• "x" = 'double quotes "inside" single quotes'
• "x" = 'single quotes 'inside' double quotes"
• "x" = 'uniçode ís tøtålly fiñe'

Booleans:

• "x" = true
• "x" = false

Arrays:

• "x" in (12, 13, 14)
• "x" in ("red", "green", "blue")

timestamp

• timestamp(YYYY-MM-DD) or timestamp(YYYY-MM-DD HH:MM) this is a helper function which generates the unix timestamp corresponding to the date (and optionally, time) entered. When the daffodil is evaluated the datetime is functionally a number, but this lets you write the daffodil in a way that is easier to read and understand. If hours and minutes are entered they are interpreted as 24 hour time UTC.
• timestamp([CURRENT_DATE][-OFFSET]) - this is a timestamp at 00:00:00 am of the given CURRENT_DATE
  • where CURRENT_DATE option may be one of the following
    • CURRENT_DAY
    • CURRENT_WEEK (week starting on Monday)
    • CURRENT_MONTH
    • CURRENT_YEAR
  • and OFFSET may be any positive integer, so expressions like the following are legit:
    • CURRENT_DAY-3 (3 days before today 00:00:00)
    • CURRENT_WEEK-10 (10 weeks before the beginning of this week)
    • ...

Examples:

• people who began mystudy after halloween 2017: mystudy__started >= timestamp(2017-10-31)

• people who participated in balloonstudy while the Macy's thanksgiving day parade was on tv (Nov 23, 9AM to 12PM Eastern time):

  balloonstudy__started >= timestamp(2017-11-23 2:00)
  balloonstudy__started < timestamp(2017-11-23 17:00)
  

• people who participated in balloonstudy this month, meaning "balloonstudy__started timestamp greater than the timestamp taken at 00:00:00 am on the first day of the current month":

  balloonstudy__started > timestamp(CURRENT_MONTH)
  

  >= and = can be used interchangeably. The only difference is that >= will treat 00:00:00 am of the first day as a timestamp belonging to the CURRENT_MONTH, while > would treat it as a timestamp belonging to the last day of the previous month.

• people who participated in balloonstudy this month:

  balloonstudy__started > timestamp(CURRENT_MONTH)
  

  Note: If > gets replaced with = it would match only those who participated exactly at 00:00:00 am (early morning today), which is somethign you probably don't want.

• people who participated in balloonstudy this week except today:

  balloonstudy__started > timestamp(CURRENT_WEEK)
  balloonstudy__started < timestamp(CURRENT_DAY)
  

• people who participated in balloonstudy during the past 5 months:

  balloonstudy__started > timestamp(CURRENT_MONTH-5)
  

Comments begin with a hash mark (#) and continue to the end of the line, similar to Python comments synthax. Generally, comments look something like this:

# this is a comment
mp_birth_year = 2002

In a more advanced example below comments come in different positions, some of them being inline, some multiline:

"mp_ethn_hispanic - not_hispanic_origin" = "no"
[
 # comment 1 (standard)
 {
  extend_fam_relate1 = "grandson"
  extend_fam_age1 = "6_11yrs" # Comment 2 (inline)
 }
{
  #
  # Comment 3 (this one is multiline)
  #
  extend_fam_relate1 = "grand_daughter"
  extend_fam_age1 = "6_11yrs"
 }
]

One may find handy the ability of commenting out an expressions checking out how expressions performs with part of it being temporarily excluded:

[
reliant_2_q_zu_income = "100k_124k"
reliant_2_q_zu_income = "125k_149k"
# this won't enter filter: reliant_2_q_zu_income = "more_150k"
]

When the entire expressions is commented out it'll match all, just like an empty expression or like {} seen above:

# {
#   age = 18
#   gender = "female"
# }