Date:         Tue, 12 Sep 2000 21:27:29 -0400
Reply-To:     paula <icj808@USWEST.NET>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         paula <icj808@USWEST.NET>
Subject:      Re: Reading four-line records

I have to admit this is not a very well put question. The problem is
resolved. This is a report used on mainframe. I simply asked the coder to
unformat the data back to one record per line. The problem is solved.

paula D
"paula" <icj808@uswest.net> wrote in message
news:GZWt5.1021$M06.243198@news.uswest.net...
> I have a text file that looks like this
>
> 1122Issuer bin: 444499    Bankruptcy alerts
date:
> 00/00/2000
> 1122
>
1122------------------------------------------------------------------------
> -----------
> 1122                                 address
> 1122trace/area                 field5
> Date89
> 1122act3/field6             field8            Date3
> field76
>
1122------------------------------------------------------------------------
> -----------
> 1122                           larkin drive
> 1122a4444w23           ACP92
00249
> 1122cancel0026          None                00266                    buy
>
1122------------------------------------------------------------------------
> ----------
>
> This is how a typical record looks like, read under WordPad under Windows
> NT.
> The top block are field names. The bottom half are readings corresponding
to
> each field on the top. All the fields as presented here are simplied. For
> example, the address field is much longer that it appears here.
>
> What I want to accomplish is to read them into SAS. I want to get rid of
the
> top block; since in SAS view, the column labels will tell what the fields
> are. I also want to rid of the header part starting with Issuer Bin....
>
> Each record looks like this.
>
> I am experimenting this way: Count from the left side of the first line
> (1122Issuer bin: 444499...)  to the last dash on the right end of the last
> line. The total is my record length. The format is Fixed, since each
record
> follows the same regularity. Then assign fields for the field names and
the
> values for the fields respectively and differently. Finally, use Keep to
> keep only the value fields.
>
> I don't know if this is going to work. I was told there will be files
coming
> in everyday. No body knows if they are similar to each other. Manuaaly
> wading through everyday is a nightmare. I wonder if anybody has better
idea
> to parse this. TIA.
>
> Paula D
>
>