Do NOT open a CSV file with EXCEL. It will change the contents on you.
That file is not data. From the header line there appear to be only three fields, but the longest lines are over 70 thousands bytes long. Also it looks like the first field is supposed to be a record number or identifier, but some of the lines (see lines 6,000 and 20,000 below) do not start with a digit string followed by a comma. So most likely there are extra line feeds inserted into some of the fields.
101 data _null_;
102 infile 'c:\downloads\Suicide_Detection.csv.zip' zip member='*' lrecl=1000000 end=eof termstr=lf;
103 input;
104 if _N_ in (1,2,3,6000,20000) or eof then list;
105 run;
NOTE: The infile 'c:\downloads\Suicide_Detection.csv.zip' is:
Filename=c:\downloads\Suicide_Detection.csv.zip,
Member Name=*
RULE: ----+----1----+----2----+----3----+----4----+----5----+----6----+----7----+----8----+----9----+----0
1 ,text,class 11
2 2,"Ex Wife Threatening SuicideRecently I left my wife for good because she has cheated on me twice a
101 nd lied to me so much that I have decided to refuse to go back to her. As of a few days ago, she beg
201 an threatening suicide. I have tirelessly spent these paat few days talking her out of it and she ke
301 eps hesitating because she wants to believe I'll come back. I know a lot of people will threaten thi
401 s in order to get their way, but what happens if she really does? What do I do and how am I supposed
501 to handle her death on my hands? I still love my wife but I cannot deal with getting cheated on aga
601 in and constantly feeling insecure. I'm worried today may be the day she does it and I hope so much
701 it doesn't happen.",suicide 727
RULE: ----+----1----+----2----+----3----+----4----+----5----+----6----+----7----+----8----+----9----+----0
3 3,Am I weird I don't get affected by compliments if it's coming from someone I know irl but I feel r
101 eally good when internet strangers do it,non-suicide 152
6000 chest pain 10
20000 0
1025022 348110,I still haven't beaten the first boss in Hollow Knight. I've only fought it a few times and I
101 always die really early in the fight. I'm terrible at this game y'all. :(,non-suicide 186
NOTE: 1025022 records were read from the infile 'c:\downloads\Suicide_Detection.csv.zip'.
The minimum record length was 0.
The maximum record length was 72599.
So the first thing to do is fix the embedded linefeeds. You can use this %replace_crlf() macro.
Then perhaps the data mining tool can read really long strings? Does it use VIYA or SAS?
filename csv temp;
%replace_crlf('c:\downloads\Suicide_Detection.csv.zip' zip member='*',csv)
If you have to read it with SAS then you will need to read the lines into multiple records of shorter strings. The maximum length of a character variable in SAS is only 32,767 bytes. So you cannot read that text into a single field.
To tease out the "TYPE" field at the end of those long lines will be a little tricky.
data suicide ;
infile csv dsd truncover length=ll firstobs=2 termstr=lf lrecl=1000000;
length recno length index 8 string $500 ;
input recno @;
length=ll ;
n=ceil(length/500);
do index=1 to n;
input string $char500. @;
output;
end;
run;
data type;
set suicide;
where index>=(n-1);
by recno;
length type $12 string2 $1000;
string2=cats(lag(string),string);
if last.recno;
type=scan(string2,-1,',');
keep recno type;
run;
data suicide;
merge type suicide;
by recno;
run;
Results:
NOTE: The infile CSV is:
Filename=C:\Users\ABERNA~1\AppData\Local\Temp\1\SAS Temporary Files\_TD12972_AMRL20L6F1E4992_\#LN00061,
RECFM=V,LRECL=1000000,
File Size (bytes)=167696307,
Last Modified=13Feb2022:23:26:56,
Create Time=13Feb2022:23:26:23
NOTE: 232074 records were read from the infile CSV.
The minimum record length was 17.
The maximum record length was 117797.
NOTE: The data set WORK.SUICIDE has 465570 observations and 5 variables.
NOTE: DATA statement used (Total process time):
real time 0.68 seconds
cpu time 0.54 seconds
104
105 data type;
106 set suicide;
107 where index>=(n-1);
108 by recno;
109 length type $12 string2 $1000;
110 string2=cats(lag(string),string);
111 if last.recno;
112 type=scan(string2,-1,',');
113 keep recno type;
114 run;
NOTE: There were 321711 observations read from the data set WORK.SUICIDE.
WHERE index>=(n-1);
NOTE: The data set WORK.TYPE has 232074 observations and 2 variables.
NOTE: DATA statement used (Total process time):
real time 0.47 seconds
cpu time 0.45 seconds
115
116 data suicide;
117 merge type suicide;
118 by recno;
119 run;
NOTE: There were 232074 observations read from the data set WORK.TYPE.
NOTE: There were 465570 observations read from the data set WORK.SUICIDE.
NOTE: The data set WORK.SUICIDE has 465570 observations and 6 variables.
NOTE: DATA statement used (Total process time):
real time 0.61 seconds
cpu time 0.50 seconds
120
121
122
123 proc freq data=suicide;
124 where index=1;
125 tables type;
126 run;
NOTE: There were 232074 observations read from the data set WORK.SUICIDE.
WHERE index=1;
NOTE: PROCEDURE FREQ used (Total process time):
real time 0.40 seconds
cpu time 0.17 seconds
The FREQ Procedure
Cumulative Cumulative
type Frequency Percent Frequency Percent
----------------------------------------------------------------
non-suicide 116037 50.00 116037 50.00
suicide 116037 50.00 232074 100.00
Now you at least have the strings in SAS variables. How to make any sense of them is up to you.
I will leave as an exercise how to remove the type from the ends of the last one or two values of STRING for each RECNO.
