Page tree

Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: Published by Scroll Versions from space DEV and version next

D toc

Excerpt

Ternary operators allow you to build if/then/else conditional logic within your transforms. Please use the IF function instead.

Info

NOTE: Ternary operators have been superseded by the IF function. See IF Function.

In the following, if the test expression evaluates to true, the true_expression is executed. Otherwise, the false_expression is executed. 

Code Block
(test_expression) ? (true_expression) : (false_expression)

All of these expressions can be constants (strings, integers, or any other supported literal value) or sophisticated elements of logical, although the test expression must evaluate to a Boolean value.

Usage

Example data:

XY
truetrue
truefalse
falsetrue
falsefalse

Transforms:

D trans
RawWrangletrue
p03Value'equals'
Typestep
WrangleTextderive type:single value:(X == Y) ? 'yes' : 'no' as: 'equals'
p01NameFormula type
p01ValueSingle row formula
p02NameFormula
p02Value(X == Y) ? 'yes' : 'no'
p03NameNew column name
SearchTermNew formula

Results:

Your output looks like the following:

XYequals
truetrueyes
truefalseno
falsetrueno
falsefalseyes

D s
snippetExamples

Example - Stock Quotes

You have a set of stock prices that you want to analyze. Based on a set of rules, you want to determine any buy, sell, or hold action to take. 

Source:

TicketQtyBuyPriceCurrentPrice
GOOG10705.25674.5
FB10084.00101.125
AAPL50125.2597.375
MSFT10038.87545.25

Transformation:

You can perform evaluations of this data using ternary operators to determine if you want to take action.

Info

NOTE: In a larger dataset, you might maintain your buy, sell, and hold evaluations for each stock in a separate dataset that you join to the source dataset before performing comparisons between column values. See Join Window.

To assist in evaluation, you might first want to create columns that contain the cost (Basis) and the current value (CurrentValue) for each stock:

D trans
RawWrangletrue
p03Value'Basis'
Typestep
WrangleTextderive type:single value:(Qty * BuyPrice) as:'Basis'
p01NameFormula type
p01ValueSingle row formula
p02NameFormula
p02Value(Qty * BuyPrice)
p03NameNew column name
SearchTermNew formula

D trans
RawWrangletrue
p03Value'CurrentValue'
Typestep
WrangleTextderive type:single value:(Qty * CurrentPrice) as:'CurrentValue'
p01NameFormula type
p01ValueSingle row formula
p02NameFormula
p02Value(Qty * CurrentPrice)
p03NameNew column name
SearchTermNew formula

Now, you can build some rules based on the spread between Basis and CurrentValue.

The most important action is determining if it is time to sell. The following rule writes a sell notification if the current value is $1000 or more than the cost. Otherwise, no value is written to the action column.

D trans
RawWrangletrue
p03Value'action'
Typestep
WrangleTextderive type:single value:(CurrentValue - 1000 > Basis) ? 'sell' : '' as:'action'
p01NameFormula type
p01ValueSingle row formula
p02NameFormula
p02Value(CurrentValue - 1000 > Basis) ? 'sell' : ''
p03NameNew column name
SearchTermNew formula

But what about buying more? The following transform is an edit to the previous one. In this new version, the sell test is performed, and if writes a buy action if the CurrentPrice is within 10% of the BuyPrice

This second evaluation is performed after the first one, as it replaces the else clause, which did nothing in the previous version. In the Recipe panel, click the previous transform and edit it, replacing it with the new version:

D trans
RawWrangletrue
p03Value'action'
Typestep
WrangleTextderive type:single value: ((CurrentValue - 1000) > Basis) ? 'sell' : ((abs(CurrentValue - Basis) <= (Basis * 0.1)) ? 'buy' : 'hold') as: 'action'
p01NameFormula type
p01ValueSingle row formula
p02NameFormula
p02Value((CurrentValue - 1000) > Basis) ? 'sell' : ((abs(CurrentValue - Basis) <= (Basis * 0.1)) ? 'buy' : 'hold')
p03NameNew column name
SearchTermNew formula

If neither test evaluates to true, the written action is hold

You might want to format some of your columns using dollar formatting, as in the following:

Info

NOTE: The following formatting inserts a dollar sign ($) in front of the value, which changes the data type to String.

D trans
RawWrangletrue
Typestep
WrangleTextset col:BuyPrice value: NUMFORMAT(BuyPrice, '$ ##,###.00')
p01NameColumns
p01ValueBuyPrice
p02NameFormula
p02ValueNUMFORMAT(BuyPrice, '$ ##,###.00')
SearchTermEdit column with formula

Results:

After moving your columns, your dataset should look like the following, if you completed the number formatting steps:

TicketQtyBuyPriceCurrentPriceBasisCurrentValueaction
GOOG10705.25$ 674.50$ 7,052.50$ 6,745.00buy
FB10084.00$ 101.13$ 8,400.00$ 10,112.50sell
AAPL50125.25$ 97.38$ 6,262.50$ 4,868.75hold
MSFT10038.88$ 45.25$ 3,887.50$ 4,525.00hold

D s also
labelother