How to Match Salesforce Fields via REGEX in JavaScript?-CodePudding

Consider the following multi-line string:

CreatedDate
Account.Example_1__c
Account.LastModifiedDate
Test__r.Example_2__c
Test__r.OwnerId
Test__r.Owner.Name
Test__r.Owner.Custom__c
Test__r.Owner.Custom__r.Id

$Action.Account.New
$ObjectType.Account
$Api.Session_ID
$Label.firstrun_helptext
.
...

How can we match the Salesforce Fields and skip the Global Variables (beginning with '$') using REGEX in JavaScript?

The REGEX should only match the following:

CreatedDate
Account.Example_1__c
Account.LastModifiedDate
Test__r.Example_2__c
Test__r.OwnerId
Test__r.Owner.Name
Test__r.Owner.Custom__c
Test__r.Owner.Custom__r.Id

/[\w.] /g matches the Salesforce Fields, but it also includes the single dots and the Global Variables in the results.

It should not include ., .., ..., etc., or the Global Variables in the matches.

Additional Examples:

1) Note that this can be a single or multi-line string, and the fields can appear before and/or after other data on the same line:

For example:

Test__r.Example_1__c >>>> (Test__r.Example_2__c) <<<< $Action.Account.New >>>> ... Test__r.Example_3__c

... should match:

Test__r.Example_1__c
Test__r.Example_2__c
Test__r.Example_3__c

2) These fields are used in formulas (like Excel formulas), so the following:

Example_1__c/Example_2__c*Example_3__c-Example_4__c Example_5__c<>Example_6__c!=Example_7__c,Example_8__c

... should return:

Example_1__c
Example_2__c
Example_3__c
Example_4__c
Example_5__c
Example_6__c
Example_7__c
Example_8__c

CodePudding user response：

How about this one?

(?<!$)(?<=^|\n|\s|\()\w [\w\.]*

We ensure that the section does not start with "$", then mark the start as either the beginning of the string, anything that comes after newline or space, or after an opening parentheses as a special case. It then matches any word that does not start with a ".", then allows dots.

Try it.

CodePudding user response：

Match the $ word/dot chars to skip these matches, and match and capture the rest with word/dot pattern:

/\$[\w.] |(\w (?:\.\w )*)/g

See the regex demo. Details:

\$[\w.] - $ and then one or more word/dot chars
| - or
(\w (?:\.\w )*) - Group 1: one or more word chars, and then zero or more sequences of a dot and then one or more word chars.

See the JS demo:

const text = 'Test__r.Example_1__c >>>> (Test__r.Example_2__c) <<<< $Action.Account.New >>>> ... Test__r.Example_3__c';
const rx = /\$[\w.] |(\w (?:\.\w )*)/g;
const matches = Array.from([...text.matchAll(rx)], (x) => x[1]).filter(Boolean);
console.log(matches);

The .filter(Boolean) removes undefined values from the result that appear because of the current pattern logic where we exclude matches without a capture.