Table row validation and column validation (r9)

The row validate and column validate tags (alias row validation and column validation) are extensions of the validate tag for use in tables. Use row or column validations to apply a validation individually to every row or column of a table.

Syntax

As with standard validations, row and column validations consist of a condition followed by a message decorator. 

Both the condition and the message support the use of option data such as [id] or [text]. The validation will be iterated across the row or column list, and the ID or text (or other option data) will be substituted. This is the key feature that distinguishes these tags from standard validations. Row and column validations can only be added to tables – the question must have a row or column list for the validation to iterate over.

row validate: [condition] {message: message text if condition not met}

column validate: [condition] {message: message text if condition not met}

Multi-line validation

Row and column validations may also be written with the message on a separate line below the validate condition. Use the row validate condition (alias rowval) and row validate message (alias rowmsg) tags for this format. Column validations work in a similar manner, instead using column validate condition (alias colval) and column validate message (alias colmsg).

# separate line column validation conditions and messages
col validate condition: [condition]
col validate message: [message]

# using colval/colmsg aliases
colval: [condition]
colmsg: [message]

# separate line row validation conditions and messages
row validate condition: [condition]
row validate message: [message]

# using rowval/rowmsg aliases
rowval: [condition]
rowmsg: [message]

This format allows for the use of Here documents (heredoc) so that complex conditions can be written on multiple lines, which helps with code readability. See the example below.

Examples

No more than two options checked per row

The following example allows respondents to select up to two (2) options per row. If the respondent selects more than 2, the row validation condition will be violated, and the validation message will be shown, prompting the respondent to "not select more than 2 attributes" for a given row element. Note the row validation message contains [text] to pipe the specific row option text into the displayed message(s).

See demo See code Restart demo

10. Please select up to two attributes you feel most exemplify each brand.
type: checkbox table
row validate: countChecked($Q10R[id]) <= 2 { message: Please do not select more than 2 attributes for [text]. }
1. Texture
2. Saltiness
3. Nutty flavor
4. Consistency
5. Aroma
rows: 
1. Skippy
2. Jif
3. Peter Pan
4. Planters
5. Store brand

No more than two options checked per column

The mechanics for a column validation work in a similar fashion, but in the context of vscale ("vertical") tables. In such cases, the table generates data points by column, rather than by row. The column validation allows SPs to check a condition against each one of the column's selections. 

See demo See code Restart demo

11. Please select up to two attributes you feel most exemplify each brand.
type: checkbox table
vscale: y
col validate: countChecked($Q11C[id]) <= 2 { message: Please do not select more than 2 attributes for [text]. }
1. Texture
2. Saltiness
3. Nutty flavor
4. Consistency
5. Aroma
cols: 
1. Skippy
2. Jif
3. Peter Pan
4. Planters
5. Store brand

Multicolumn table

In this example, respondents must give an assessment of a brand in column B if they are familiar with it in column A. The enableWhen tag is included on column B so that it is only available when respondents indicate familiarity with one of the brands (selecting options 2-5) in column A.

Note that because this is a multicolumn table, start table / end table syntax is needed instead of simple table syntax.

See demo See code Restart demo

set list: BRANDS
1. Skippy
2. Jif
3. Peter Pan
4. Planters
5. Store brand

start table: 12
intro: For each brand you're familiar with, please rate your assessment on a scale of 1-5.
rows from: BRANDS
row validate: !(anyChecked($Q12R[id]CA,2..5) and noneChecked($Q12R[id]CB,1..4)) { message: Please give an assessment of [text] since you are familiar with it. }
--
column id: A
Type: radio
heading: Awareness
1. Not at all familiar
2. Slightly familiar
3. Somewhat familiar
4. Very familiar
5. Extremely familiar
--
column id: B
enablewhen: anyChecked($Q12R__RID__CA,2..5)
Type: radio
heading: Assessment
required: n
1. Awful
2. Mediocre
3. Good
4. Exceptional
end table

Applying multiple row validations

Multiple row validations may be applied to a table by defining them in a heredoc within the row validations tag. Line spaces are required between each set of conditions and messages within the heredoc since the tags are repeated.

In this example, the first two validations ensure that a text response is entered in the column corresponding to the radio option selected. Then, the third and fourth validations ensure that a text response is not entered in the column that doesn't match the radio option selected.

See demo See code Restart demo

set list: GROCERYSTORES
1. Trader Joes
2. Costco
3. Publix

start table: 13
intro: For each store, indicate if you shop there. If you do, please specify the reason for shopping. If you don't, specify the reason you aren't a customer.
row validate: not (anyChecked($Q13R[id]CA,1) and $Q13R[id]CB eq '') { message: For stores you DO shop at, please specify your reason in the "why" column. }
row validate: not (anyChecked($Q13R[id]CA,2) and $Q13R[id]CC eq '') { message: For stores you DO NOT shop at, please specify your reason in the "why not" column. }
row validate: not (anyChecked($Q13R[id]CA,1) and $Q13R[id]CC ne '') { message: For stores you DO shop at, please do not specify a reason in the "why not" column. }
row validate: not (anyChecked($Q13R[id]CA,2) and $Q13R[id]CB ne '') { message: For stores you DO NOT shop at, please do not specify your reason in the "why" column. }
rowsfrom: GROCERYSTORES
--
column id: A
type: radio
1. Do shop there
2. Never shop there
required: y
--
column id: B
type: text
size: 15
required: n
heading: Why do you shop there?
--
column id: C
type: text
size: 15
required: n
heading: Why don't you shop there?
end table

Heredoc condition example

In some situations, validation conditions can be lengthy. To improve code readability, use a multi-line row or column validation with a heredoc at the condition. This allows for line breaks within the condition itself.

In the example below, the numeric responses in Q16 are validated to check that they fit in the range selected at Q15, requiring a lengthy condition. A heredoc is used to separate the condition onto multiple lines, based on the expressions between each or operator.

See demo See code Restart demo

setlist: STORES 
1. Flatley and Sons
2. Brandi's

15. How many times did you shop at the following retailers last year?
type: radio table 
rowsfrom: STORES
1. 1-5
2. 6-10
3. 11+

16. To the best of your knowledge, what is the specific number of times you shopped at each retailer last year?
type: number table
ap: n
maxlen: 2 
rowsfrom: STORES
row validate condition: <<END 
(anyChecked($Q15R[id],1) and $Q16R[id] >= 1 and $Q16R[id] <= 5) or
(anyChecked($Q15R[id],2) and $Q16R[id] >= 6 and $Q16R[id] <= 10) or
(anyChecked($Q15R[id],3) and $Q16R[id] >= 11) 
END 
row validate message: You said you shopped at [text] $Q15R[id]_text times. Please make sure your specific number is within that range.