How to Parse a VBW Message Using Regular Expressions¶

1. Example VBW Sentence¶

We'll start with a real VBW message:

$VDVBW,8.2,0.5,A,8.1,0.4,A,8.3,A,8.2,A*57

VBW is an NMEA-style sentence consisting of:

A start character ($)
A sentence identifier (VDVBW)
A sequence of comma-separated fields
A * followed by a two-digit hexadecimal checksum

Every field contains either a value or may be blank.

2. Understanding the Fields¶

A VBW message contains 10 comma-separated data fields after the sentence ID:

Field	Meaning (simplified)
1	Longitudinal water speed
2	Transverse water speed
3	Status (water speed)
4	Longitudinal ground speed
5	Transverse ground speed
6	Status (ground speed)
7	Transverse water speed (stern sensor)
8	Status (stern water speed)
9	Transverse ground speed (stern sensor)
10	Status (stern ground speed)

A field may be present:

8.2

or blank:

So our regex needs to allow for both situations.

3. Start Building the Regex¶

Step 1 – Match the sentence start¶

^\$VDVBW,

^ = start of the line
\$ = literal $
VDVBW, = the sentence identifier and the first comma

Step 2 – Capturing Fields¶

Since any field may be blank, the easiest and most robust pattern is:

[^,]*

Meaning:

"Match zero or more characters until the next comma."

[^,] = match any character except a comma
* = zero or more times

Note about character classes: The square brackets [] create a character class that matches any single character from the set. When you add ^ inside the brackets (like [^,]), it creates a negated character class that matches any character NOT in the set. So [^,] matches letters, numbers, periods, spaces - anything except a comma.

To store the value, we put this into a named capture group:

(?P<long_w_spd>[^,]*)

We repeat this pattern for each of the ten fields.

Step 3 – Add the Checksum¶

VBW ends with:

*HH

Where HH is a two-digit hexadecimal checksum. That can be matched with:

\*[0-9A-Fa-f]{2}

4. Final Complete Regex¶

This expression:

Matches a valid VBW sentence
Captures every field into a named variable
Allows any field to be blank
Requires a valid 2-digit hexadecimal checksum

Single-line version¶

^\$VDVBW,(?P<long_w_spd>[^,]*),(?P<tran_w_spd>[^,]*),(?P<stat_w_spd>[^,]*),(?P<long_g_spd>[^,]*),(?P<tran_g_spd>[^,]*),(?P<stat_g_spd>[^,]*),(?P<tran_w_spd_stern>[^,]*),(?P<stat_w_spd_stern>[^,]*),(?P<tran_g_spd_stern>[^,]*),(?P<stat_g_spd_stern>[^,]*)\*[0-9A-Fa-f]{2}$

Readable multi-line version (if your regex flavor supports verbose mode)¶

^
\$VDVBW,
(?P<long_w_spd>[^,]*),
(?P<tran_w_spd>[^,]*),
(?P<stat_w_spd>[^,]*),
(?P<long_g_spd>[^,]*),
(?P<tran_g_spd>[^,]*),
(?P<stat_g_spd>[^,]*),
(?P<tran_w_spd_stern>[^,]*),
(?P<stat_w_spd_stern>[^,]*),
(?P<tran_g_spd_stern>[^,]*),
(?P<stat_g_spd_stern>[^,]*)
\*[0-9A-Fa-f]{2}
$

Once matched, each field is available by its name. For example:

long_w_spd
tran_w_spd
stat_w_spd, etc.

You can then convert, validate, or interpret them however your application requires.

5. Using the Results in CORIOLIX¶

CORIOLIX expects a Python list of regex strings.

['regex1']

For a single message type like VBW, you provide:

['^\$VDVBW,(?P<long_w_spd>[^,]*),(?P<tran_w_spd>[^,]*),(?P<stat_w_spd>[^,]*),(?P<long_g_spd>[^,]*),(?P<tran_g_spd>[^,]*),(?P<stat_g_spd>[^,]*),(?P<tran_w_spd_stern>[^,]*),(?P<stat_w_spd_stern>[^,]*),(?P<tran_g_spd_stern>[^,]*),(?P<stat_g_spd_stern>[^,]*)\\*[0-9A-Fa-f]{2}$']

If a sensor outputs more than one message type (like VBW and DPT), you provide one regex per type:

['regex1', 'regex2', ...]