Fix the BAI2 grammar from forked package #1
Conversation
| @@ -90,7 +90,7 @@ | |||
| "detailFundsType": "", | |||
| "bankReferenceNumber": "AGN", | |||
| "customerReferenceNumber": "935892", | |||
| "detailText": "POINT OF SALE" | |||
| "detailText": "POINT OF SALE/" | |||
There was a problem hiding this comment.
I make a few changes to the test cases, because I believe the grammar was implemented incorrectly.
The specification states that
Therefore, the end of a record ending in text cannot be delimited using these characters. Instead, the logical record continues until another record that is not a continuation (88) record begins
| senderIdentification = alnum + | ||
| receiverIdentification = alnum + | ||
| senderIdentification = alphaNumeric+ | ||
| receiverIdentification = alphaNumeric+ |
There was a problem hiding this comment.
These alterations are hardening the definition, these values cannot include , or /
| @@ -74,7 +74,7 @@ BTR3 { | |||
|
|
|||
| amountOpt = optSign digit* | |||
| itemCount = digit* | |||
| fundsType = alnum? | |||
| fundsType = ("0".."2" | "S" | "V" | "D" | "Z") ? | |||
There was a problem hiding this comment.
Again, hardening the definition, there is a specific set of values for this [pg. 29]
| @@ -102,8 +102,10 @@ BTR3 { | |||
| bankReferenceNumber = alphaNumeric* | |||
| customerReferenceNumber = alphaNumeric* | |||
| detailText (Text relaxed for CBA BAI2 format does not contain , / Optionally followed by /) | |||
| = (~ eor any) * | |||
| = & "/" // null text field | |||
| | text * // normal text field | |||
There was a problem hiding this comment.
Here, we need to do a lookahead for a / (in the file, this would look like ,/ because detailText is preceded by delim. In this case, the text is null, and trying to parse it as a text would cause the expression to be unreasonably complex
| @@ -120,10 +122,17 @@ BTR3 { | |||
| hh = ("0".."1" digit) | ("2" "0" .. "3") | |||
| mm = "0".."5" digit | |||
|
|
|||
| delim = "," | |||
| newLine = ("\r\n" | "\n" | "\r") | |||
There was a problem hiding this comment.
We should accept all kinds of newlines
| delim = | ||
| | "," -- comma_delimiter | ||
| | "/" newLine "88," -- continuation_record_delimiter | ||
|
|
There was a problem hiding this comment.
This is big. I have added support for 88 - Continuation Record [pg. 20]. It is simplest to treat \n88, as a delimiter rather than a new record.
There was a problem hiding this comment.
If a record is "split" between fields, it will look like
...field1,field2/
88,field3,field4...
which we want to parse as field1,field2,field3,field4
The slash replaces the comma that ordinarily delimits that field. The example below correctly illustrates the use of delimiters.
| text = (continuationWithinTextField | ~newLine any) |
There was a problem hiding this comment.
However, a text field does not have a delimiter at the end of a logical record. we want to parse
...text
88,andmoretext
as textandmoretext
Here is the specification for the file format
Cash Management Balance Reporting Specifications Version 2.pdf
I make a few changes to this code: