PRDateFormat.parseAsDateTime is not working as expected
As per Engine API definition,
public static final java.util.Date parseAsDateTime(java.lang.String aLocale,
java.lang.String aZone,
java.lang.String[] aPatterns,
java.lang.String aValue)
convenience method to parse a String as a DateTime, trying a list of patterns in an attempt to guess the correct format.
- Parameters:
aLocale- locale to useaZone- time zone to defaultaPatterns- array of patterns, if null or empty, a default list will be usedaValue- String to parse
----------------
I'm setting Patterns as below :
java.lang.String[] aPatterns = { PRDateFormat.DEFAULT_DATETIME_DEFAULT, PRDateFormat.DEFAULT_DATETIME_SHORT, PRDateFormat.DEFAULT_DATETIME_LONG, PRDateFormat.DEFAULT_DATETIME_FULL, PRDateFormat.DEFAULT_DATETIME_SHORT_YY, PRDateFormat.DEFAULT_DATETIME_SHORT_YYYY };
At Runtime can you explain which Pattern will be used ?
I want to understand how this function picks the Pattern from this array ?
***Edited by Moderator: Pallavi to update platform capability tags***
Pegasystems Inc.
CA
This is a great question thank you for posting it, @MAHESHHR
Before answering this question, it’s key to note that the answer may be dependent on a few factors:
- The input String
- The locale of the user
- The JVM used
Therefore, it’s important to ensure that your regression test suite regularly tests input and outputs to ensure compliance with business requirements.
Let’s dive into this.
First, the input String array that you mentioned above, “aPatterns”, as per the API (https://community.pega.com/sites/default/files/help_v84/javadocs/index.html), states the following
public static final java.lang.String DEFAULT_DATETIME_DEFAULT
locale-specific default datetime format
Therefore this could theoretically act differently in different locales.
Secondly, if the user provides different Strings, then the output may differ. In the case where two patterners both match the input, then most likely it will choose the first. However if you’re running a custom JVM which iterates over the array in a non-deterministic fashion, then the results may vary.
This is a great question thank you for posting it, @MAHESHHR
Before answering this question, it’s key to note that the answer may be dependent on a few factors:
- The input String
- The locale of the user
- The JVM used
Therefore, it’s important to ensure that your regression test suite regularly tests input and outputs to ensure compliance with business requirements.
Let’s dive into this.
First, the input String array that you mentioned above, “aPatterns”, as per the API (https://community.pega.com/sites/default/files/help_v84/javadocs/index.html), states the following
public static final java.lang.String DEFAULT_DATETIME_DEFAULT
locale-specific default datetime format
Therefore this could theoretically act differently in different locales.
Secondly, if the user provides different Strings, then the output may differ. In the case where two patterners both match the input, then most likely it will choose the first. However if you’re running a custom JVM which iterates over the array in a non-deterministic fashion, then the results may vary.
Therefore, if you’d like deterministic results, your best bet is to use non-locale-based-patterns where one and only one pattern will match an input string; this must be part of the business requirements.
For example, the following shows how we would accomplish just that:
java.lang.String[] aPatterns = { "yyyy.MM.dd G 'at' HH:mm:ss z" };
java.util.Date myDateObject = PRDateFormat.parseAsDateTime("en_CA","EDT",aPatterns,"2001.07.04 AD at 12:08:56 PDT");
oLog.error("myDateObject::: " + myDateObject.toString());
Output:
myDateObject::: Wed Jul 04 15:08:56 EDT 2001
In summary, the solution is to both ensure the business requirements are clear and the CICD regression test suite contains thorough test cases.
Hope this answers your question. For additional insight please see the following: