Inserting datalines into file

[Autotelic]

How does one go about adding datalines to an existing text file?

I'm on SAS 9.4, SAS EG 7.1

I tried the following (copied from somewhere), but I can't adapt it to what I need.

data _null_;
   file "file_path_name" mod;
  input fruit :$10.;
  put fruit;
datalines;
date
elderberry
fig
;
run;

I want to append semicolon separated values both to the start and to the end of the file.

As an example assume that file_path_name looks like this:

1;2;a
2;2;b

And I want to add 

0;;

to the top of the file and 

2;3;c

to the end of the file.

These lines should be present in code and not in another file.

How can I do this?

[RW9]

You cant really mod a file to add text at the beginning, you read it then write it out, something like:

data _null_;
  set yourdata;
  infile "your_file";
  file "your_file";
  /* Put first update */
  if _n_=1 then do;
    put xyz;
  end;
  input;
  put _input_;
  /* Put end update */
  if eof then do;
    put ...;
  end;
run;

I suppose the question would be why would you need to , you can just append your data, then the reading program would sort it anyways.

[Autotelic]

Actually what I want is to export a dataset as a semicolon separated value text file, only

1. on top of the dataset headers, I need to add two semicolon separated row of values, the first one is a row of headers, the second one is a row of values for the first header.
2. the bottom of the csv has the same structure as the dataset and are datalines which I want to be able to customize in code

I then want to output the result of this as a text file.

Does this way of phrasing things make it any simpler?

[RW9]

Well, firstly, and most important, delimiting a file with a semicolon and then calling it a csv doesn't work.  CSV stands for Comma Separated Variable file.

For your question, simplest way would be to read in the data as it stands, write (or copy from a proc import in the log) a datastep import (using infile) of the data as it stands.  Then with your dataset add the extra row(s) you need to the dataset - this means you can use all the SAS formats and such like.  Then once you have your final data, output this to delimited file (*.txt), using a datastep with a file command.

[Tom]

---

@Autotelic wrote:

Actually what I want is to export a dataset as a semicolon separated value text file, only

1. on top of the dataset headers, I need to add two semicolon separated row of values, the first one is a row of headers, the second one is a row of values for the first header.
2. the bottom of the csv has the same structure as the dataset and are datalines which I want to be able to customize in code

I then want to output the result of this as a text file.

Does this way of phrasing things make it any simpler?

---

It is trivial to write a delimited file using a data step. Just use the DSD= and DLM= option on the FILE statement and write the data.

file 'myfile.txt' dsd dlm=';' ;
put a b c ;

If you want to write extra information at the top then add and IF _N_=1 THEN DO;... END; block to the data step.

if _n_=1 then do;
  put 'header1';
  put x1 x2 x3 ;
end;

Does the "bottom" of the file include a new header row?

Please post an example dataset (in the form of a data step that we can run to create the example data) .  Also include source for the information you want to put in the first two lines.  Then show an example of the output you want to create from that sample data.