Greetings,

There have been a number of messages in this group lately
about reading comma delimited *.csv files in J.  This is something
I've been doing for years with the following verb.  It works
quite well for small and modestly sized *.csv files.  Large files will
have to be processed in chunks.

NB. J script begins *********************************************

CR=: 13 { a.

NB. assertion error with message
assertmsg=: [ 13!:8 ([: 0&e. ]) $ 12"_

NB. parse comma delimited *.csv files the y. argument
NB. is a comma delimited character list.  The x. argument
NB. specifies alternate delimiters. Assumes LF , CRLF or LFCR
NB. delimited lines

parsecsv=: 3 : 0
',' parsecsv y.
:
'separater cannot be the " character' assertmsg -. x. -: '"'

NB. CRLF delimited *.csv text to char table
y. =.  x. ,. ];._2 y. -. CR

NB. bit mask of unquoted " field delimiters
b  =. -. }. ~:/\ '"' e.~  ' ' , , y.
b  =. ($y.) $ b *. , x. = y.

NB. use masks to cut lines
b <;._1"1 y.
)

NB. J script ends ***********************************************

To use this verb do the following:

   NB. read comma delimited text file and parse
   x =.  parsecsv 1!:1 <'c:\talks\rasc\mess2.csv'

   NB. corner elements of parsed file
   10 5 {. x
+-------------+--------+--------+--------+--------+
|<SEEN>       |OBJECT_I|OBJ_ALT_|CONS_ABB|OBJ_TYPE|
+-------------+--------+--------+--------+--------+
|3/27/95 21:25|M1      |NGC1952 |Tau     |PN      | <-- wrong supernova
+-------------+--------+--------+--------+--------+     remnant, never
|5/25/96 1:30 |M10     |NGC6254 |Oph     |GC      |     noticed this before
+-------------+--------+--------+--------+--------+
|             |M100    |NGC4321 |Com     |SG      |
+-------------+--------+--------+--------+--------+
|7/21/96 0:30 |M101    |NGC5457 |UMa     |SG      |
+-------------+--------+--------+--------+--------+
|7/21/96 1:00 |M102    |NGC5457 |UMa     |SG      |
+-------------+--------+--------+--------+--------+
|9/20/96 0:30 |M103    |NGC581  |Cas     |OC      |
+-------------+--------+--------+--------+--------+
|3/25/95 23:40|M104    |NGC4594 |Vir     |SG      |
+-------------+--------+--------+--------+--------+
|             |M105    |NGC3379 |Leo     |EG      |
+-------------+--------+--------+--------+--------+
|             |M106    |NGC4258 |CVn     |SG      |
+-------------+--------+--------+--------+--------+

   NB. top of raw data looks like:

<SEEN>,OBJECT_I,OBJ_ALT_,CONS_ABB,OBJ_TYPE,OBJ_RA,OBJ_DEC,OBJ_MAGV,OBJ_SIZE,
OBJ_BURN,OBJ_COMM,SITE,OPTICS
3/27/95 21:25,M1,NGC1952,Tau,PN,534.5,22.01,8.2,6x4,!!,SNR (1054) - Crab
Nebula,Home,125a/18e telescope
5/25/96 1:30,M10,NGC6254,Oph,GC,1657.1,-4.06,6.6,8.2,!,VII,Home,7*50 binoculars
,M100,NGC4321,Com,SG,1222.9,15.49,9.4,5.3x4.5,!,Sc - fine spiral,,

    ....

Hope this helps.
------------------------------------------------------------------------
"Natural selection: the ultimate focus group!
------------------------------------------------------------------------
John D. Baker