Import Filter with variable tag length

Reference Manager Reference Manager Styles, Filters, and Connections
#1

I am trying to create an import filter for a data export from my data-driven website containing medical conference records.  It is modeled on NLM MEDLINE tags.  Specification for Tag Format seems to require a fixed number of characters (e.g. \u\u–) but my exported text file has anywhere from 2 to 4 upper case characters (e.g tags include AU, OWN, PMID; 2, 3 and 4 characters).  Therefore, setting the tag format to only 2 upper case characters causes merging of data from multiple fields (e.g. AU and FAU both go into the Author field causing replication of data).  What technique can be used to permit the tags to vary in length?

#2

Hello pcherry,

When you are editing the filter click on the “Reference and Tag Format” tab.  Then set the tag format to “¶\u\u\d\d-”

This will tell RM that tags will start at the beginning of the line, then 2 uppercase characters, and 2 letters, numbers, or spaces followed by a dash. 

#3

Peter Travis,

Thanks for your reply.  Ihave attached my initial cap file.  When I make your suggested modification in the filter, nothing imports.  I will attach a sample datafile in another reply ( I seem to be able to attach only one file per message here.

Thanks,

Pete Cherry

MIS MEDLINE (Include UI).cap (900 Bytes)

#4

Hello again,

The settings I posted are if you have tags in the format of:

XX  - <Data>

XXXX- <Data>

Where you have 2-4 characters followed by a dash.

The tags in your style are in the format:

XX

There are no spaces or dashes.  If the data in the text file is "XX  - " or “XXXX-”  you need the spaces and dashes entered in the tags.

#5

last attempt at posting sample text file failed so i added os to the txt extension

rmisExportanast.os (30.7 KB)

#6

 \u\u\··-· is how i originally defined tag format.

¶¬\u\u\d\d··-· is based on your suggestion and with my text file nothing is imported.

Thank you for your patience and and for your help in advance.

#7

The file is not tagged consistantly.  If you are generating the file you want 4 digits then the dash.

The Medline format for the Full Author tag is:

FAU -  (one space between the U and -)

In your file you have:

FAU  - (two spaces between the U and -) 

#8

  1. Remove the ¬ and spaces

  2. Edit your tags so they are complete with spaces and dashes

Message Edited by PTravis on 04-14-2008 10:08 AM

#9

Peter,

Much better results.  Thanks for putting up with this naive user.

Pete Cherry

#10

Hi Peter,

The sequence ¶\u\u\d\d- works for two or four letter tags (PMID or DA), but no data is imported for three letter tags (e.g. in my text file OWN or FAU).  Please let me know if this is feasible.  I really appreciate the help.

Pete Cherry

1 Like
#11

One of the tricks I use is the “Custom Character Set”.

To use this, basically you simply use Brackets [] to set the characters that can appear in those positions.

As such, if you define the Tag Format as something like this…

\u\u[A1·][b2·]-

This will effectively allow you to create a group of possible characters, symbols, spaces to appear in those positions.

Per my example (with * indicating the group of Any Uppercase Letters, and · being a space), the resulting possibilities that the example would allow for would be…

**··-

**A2-

**1b-

**Ab-

**12-

**1b-

**·2-

**·b-

**A·-

**1·-

Using this allows for flexibility that one simply cannot get with all inclusive groups, like the \d, since you can effectively limit what Reference Manager considers to be a viable element in a given tag position.

I do hope this helps.

Thanks!

SteveH
Customer Technical Support Group Leader
Global Customer Support  Thomson Scientific
 425 Market Street
San Francisco, CA  94105
USA
Tel        +1 408 987 5609

http://www.researchsoft.com
http://scientific.thomson.com

Thomson and its product names and acronyms used herein are trademarks, service marks, and registered trademarks used under license.

1 Like