hl7api-devel

RE: [HAPI-devel] New HAPI user

[Tripp, B. <Bry...@uh...>]

Hi Vishal, 

Messages/segments/fields in an oracle database. Does that make sense?  

Yes, that's a very common use of HAPI.  

server but I will probably extend Class Parser and implement my own parser
that would actually save the parsed 

I suggest implementing ca.uhn.hl7v2.app.Application instead of extending
Parser.  The TCP/IP server will parse messages for you, and send them to
your Application's processMessage(Message) method, where you can put the
code that writes to the database.  See the HL7Service javadocs for
information on registering your application with the server (it's easy --
you make a file with a line that identifies the message types and trigger
events that you handle, and the name of your application class).  

Another thing I noticed that HAPI is restricted to start with '\u000b' and
end with '\u001c' + '\r'. Most of the people I work with, they have random
message delimiters but consistent within their facility, I mean they never
follow standards 

Wow.  I guess that shouldn't be surprising, but HAPI is almost 2 years old
and you're the first one to mention this.  Maybe the practice varies
regionally.  Where are you located?  How many times have you seen
non-standard message delimiters?  

 

Bryan 



This e-mail may contain confidential and/or privileged information
for the sole use of the intended recipient. Any review or distribution
by anyone other than the person for whom it was originally intended is
strictly
prohibited. If you have received this e-mail in error, please contact the
sender and
delete all copies. Opinions, conclusions or other information contained in
this e-mail may not be that of the organization.

RE: [HAPI-devel] New HAPI user

[Tripp, B. <Bry...@uh...>]

Hi Vishal, 
 
Yes, people put messages in files in all kinds of wierd ways.  How about
scanning the file using some kind of regular expression that will find
messages amid arbitrary content.  Like this maybe:
MSH(.).*\r\n?([A-Z]{3}\1.*\r\n?)*   (I don't know if that one will work --
just trying to illustrate).   
 
Or you could hack the MLLP code to use delimiters that you set at run time.
I think I did that once.   If I remember correctly there was some trick to
using escape sequences outside of Java source (you would want to do this so
you could easily type them into a config file or on the command line using
characters you have on your keyboard).  Nope, it's not coming to me.  

Bryan 

-----Original Message-----
From: Vishal Abrol [mailto:va...@ve...]
Sent: September 18, 2003 6:18 AM
To: 'Tripp, Bryan'
Subject: RE: [HAPI-devel] New HAPI user



Yes, messages in the file. Whenever we develop interface for somebody,
usually they send us the files as a specification.

 

Would that be different?

Vishal

 

-----Original Message-----
From: Tripp, Bryan [mailto:Bry...@uh...] 
Sent: Thursday, September 18, 2003 10:56 PM
To: 'Vishal Abrol'
Subject: RE: [HAPI-devel] New HAPI user

 

Hi Vishal, 

Vishal Answer: Really? I am actually surprised. Since yesterday I collected
six sets of HL messages from some from our customers ( real time data,5-10
messages) and actually one of the files have standard start message
character   '\u000b'  but most of the time message end  varies with  '0D 1C
OA' ,'0D 0A 1C 0D 0A',atleast '1C' is usually consistent . One of the sets
had only '0D' as start and end with '0D' again. I have been dealing with
customer all over US and yes, we did have problems with our HL7 parsing
usually ended up in custom fixes. This is one of the reasons; I need to
develop a standard solution. So far, I have been pre processing the messages
to prefix-suffix proper delimiters before I send it to my parser for
parsing. Segment delimiters are usually limited to '0A','0D','0D 0A' and '0A
0D'.I work in CA,USA and had a lot of interaction with Hospitals since 2-3
years. Hopefully you will be able to address this issue ?

Oh!  I thought you were talking about the minimal lower layer protocol.  Do
you mean messages in files?  

 

Bryan 

 



This e-mail may contain confidential and/or privileged information

for the sole use of the intended recipient. Any review or distribution

by anyone other than the person for whom it was originally intended is
strictly

prohibited. If you have received this e-mail in error, please contact the
sender and

delete all copies. Opinions, conclusions or other information contained in

this e-mail may not be that of the organization.




This e-mail may contain confidential and/or privileged information
for the sole use of the intended recipient. Any review or distribution
by anyone other than the person for whom it was originally intended is
strictly
prohibited. If you have received this e-mail in error, please contact the
sender and
delete all copies. Opinions, conclusions or other information contained in
this e-mail may not be that of the organization.

RE: [HAPI-devel] New HAPI user

[Vishal A. <va...@ve...>]

Thanks for your help.

I will try to implement this. I plan to give the user flexibility to =
define
the delimiters in a XML config file. I probably have to hack the MLLP =
code
to read those delimiters from file .will let u know once I am successful =
in
this.

Thanks

=20

-----Original Message-----
From: Tripp, Bryan [mailto:Bry...@uh...]=20
Sent: Thursday, September 18, 2003 11:33 PM
To: 'Vishal Abrol'; hl7...@li...
Subject: RE: [HAPI-devel] New HAPI user

=20

Hi Vishal,=20

=20

Yes, people put messages in files in all kinds of wierd ways.  How about
scanning the file using some kind of regular expression that will find
messages amid arbitrary content.  Like this maybe:
MSH(.).*\r\n?([A-Z]{3}\1.*\r\n?)*   (I don't know if that one will work =
--
just trying to illustrate).  =20

=20

Or you could hack the MLLP code to use delimiters that you set at run =
time.
I think I did that once.   If I remember correctly there was some trick =
to
using escape sequences outside of Java source (you would want to do this =
so
you could easily type them into a config file or on the command line =
using
characters you have on your keyboard).  Nope, it's not coming to me. =20

Bryan=20

-----Original Message-----
From: Vishal Abrol [mailto:va...@ve...]
Sent: September 18, 2003 6:18 AM
To: 'Tripp, Bryan'
Subject: RE: [HAPI-devel] New HAPI user

Yes, messages in the file. Whenever we develop interface for somebody,
usually they send us the files as a specification.

=20

Would that be different?

Vishal

=20

-----Original Message-----
From: Tripp, Bryan [mailto:Bry...@uh...]=20
Sent: Thursday, September 18, 2003 10:56 PM
To: 'Vishal Abrol'
Subject: RE: [HAPI-devel] New HAPI user

=20

Hi Vishal,=20

Vishal Answer: Really? I am actually surprised. Since yesterday I =
collected
six sets of HL messages from some from our customers ( real time =
data,5-10
messages) and actually one of the files have standard start message
character   '\u000b'  but most of the time message end  varies with  '0D =
1C
OA' ,'0D 0A 1C 0D 0A',atleast '1C' is usually consistent . One of the =
sets
had only '0D' as start and end with '0D' again. I have been dealing with
customer all over US and yes, we did have problems with our HL7 parsing
usually ended up in custom fixes. This is one of the reasons; I need to
develop a standard solution. So far, I have been pre processing the =
messages
to prefix-suffix proper delimiters before I send it to my parser for
parsing. Segment delimiters are usually limited to '0A','0D','0D 0A' and =
'0A
0D'.I work in CA,USA and had a lot of interaction with Hospitals since =
2-3
years. Hopefully you will be able to address this issue ?

Oh!  I thought you were talking about the minimal lower layer protocol.  =
Do
you mean messages in files? =20

=20

Bryan=20

=20

=20

This e-mail may contain confidential and/or privileged information

for the sole use of the intended recipient. Any review or distribution

by anyone other than the person for whom it was originally intended is
strictly

prohibited. If you have received this e-mail in error, please contact =
the
sender and

delete all copies. Opinions, conclusions or other information contained =
in

this e-mail may not be that of the organization.

=20



This e-mail may contain confidential and/or privileged information

for the sole use of the intended recipient. Any review or distribution

by anyone other than the person for whom it was originally intended is
strictly

prohibited. If you have received this e-mail in error, please contact =
the
sender and

delete all copies. Opinions, conclusions or other information contained =
in

this e-mail may not be that of the organization.

RE: [HAPI-devel] New HAPI user

[Vishal A. <va...@ve...>]

Bryan,=20

       =20

        That was a very quick response. Appreciate it.=20

       =20

server but I will probably extend Class Parser and implement my own =
parser
that would actually save the parsed=20

I suggest implementing ca.uhn.hl7v2.app.Application instead of extending
Parser.  The TCP/IP server will parse messages for you, and send them to
your Application's processMessage(Message) method, where you can put the
code that writes to the database.  See the HL7Service javadocs for
information on registering your application with the server (it's easy =
--
you make a file with a line that identifies the message types and =
trigger
events that you handle, and the name of your application class). =20

=20

Vishal Answer:  Yep, thanks, I saw that and probably will do that . Only
reason I was trying to extend Parser because of having hard time
understanding API structures and actually parsing/retrieving values from =
the
messages using HAPI API's . Do you have sample parsing routines and how =
to
access values from Hl7 messages? I noticed few in the API DOCS but still
unclear. I guess I will play around with it more.  But advantage of
extending the parser is that I have control on raw message in the class =
file
and I could just call my API to start sending the segments/fields  =
during
parsing to HAPI structures.

=20

=20

Another thing I noticed that HAPI is restricted to start with '\u000b' =
and
end with '\u001c' + '\r'. Most of the people I work with, they have =
random
message delimiters but consistent within their facility, I mean they =
never
follow standards=20

Wow.  I guess that shouldn't be surprising, but HAPI is almost 2 years =
old
and you're the first one to mention this.  Maybe the practice varies
regionally.  Where are you located?  How many times have you seen
non-standard message delimiters? =20

=20

=20

Vishal Answer: Really? I am actually surprised. Since yesterday I =
collected
six sets of HL messages from some from our customers ( real time =
data,5-10
messages) and actually one of the files have standard start message
character   '\u000b'  but most of the time message end  varies with  '0D =
1C
OA' ,'0D 0A 1C 0D 0A',atleast '1C' is usually consistent . One of the =
sets
had only '0D' as start and end with '0D' again. I have been dealing with
customer all over US and yes, we did have problems with our HL7 parsing
usually ended up in custom fixes. This is one of the reasons; I need to
develop a standard solution. So far, I have been pre processing the =
messages
to prefix-suffix proper delimiters before I send it to my parser for
parsing. Segment delimiters are usually limited to '0A','0D','0D 0A' and =
'0A
0D'.I work in CA,USA and had a lot of interaction with Hospitals since =
2-3
years. Hopefully you will be able to address this issue ?

=20

Thank you for your input.

Vishal

=20

 =20

=20

=20

=20

-----Original Message-----
From: Tripp, Bryan [mailto:Bry...@uh...]=20
Sent: Thursday, September 18, 2003 8:49 PM
To: 'Vishal Abrol'; hl7...@li...
Subject: RE: [HAPI-devel] New HAPI user

=20

Hi Vishal,=20

Messages/segments/fields in an oracle database. Does that make sense? =20

Yes, that's a very common use of HAPI. =20

server but I will probably extend Class Parser and implement my own =
parser
that would actually save the parsed=20

I suggest implementing ca.uhn.hl7v2.app.Application instead of extending
Parser.  The TCP/IP server will parse messages for you, and send them to
your Application's processMessage(Message) method, where you can put the
code that writes to the database.  See the HL7Service javadocs for
information on registering your application with the server (it's easy =
--
you make a file with a line that identifies the message types and =
trigger
events that you handle, and the name of your application class). =20

Another thing I noticed that HAPI is restricted to start with '\u000b' =
and
end with '\u001c' + '\r'. Most of the people I work with, they have =
random
message delimiters but consistent within their facility, I mean they =
never
follow standards=20

Wow.  I guess that shouldn't be surprising, but HAPI is almost 2 years =
old
and you're the first one to mention this.  Maybe the practice varies
regionally.  Where are you located?  How many times have you seen
non-standard message delimiters? =20

=20

Bryan=20



This e-mail may contain confidential and/or privileged information

for the sole use of the intended recipient. Any review or distribution

by anyone other than the person for whom it was originally intended is
strictly

prohibited. If you have received this e-mail in error, please contact =
the
sender and

delete all copies. Opinions, conclusions or other information contained =
in

this e-mail may not be that of the organization.