One alternative using non-array formulas which automates the extraction of
the uniques list from 4 source cols and the corresponding counts of the
uniques ..

Illustration in this sample construct:
http://www.savefile.com/files/841298
Multicol merge n uniques extract n count.xls

Source data assumed in cols A to F, data from row5 down
where col C = Lube data, col D = air-primary data,
col E = air-secondary data, col F = fuel-inner data

Using 4 adjacent cols to the right, cols G to J

In G5:
=OFFSET(C$5,INT((ROWS($1:1)-1)/4),MOD(ROWS($1:1)-1,4))
Copy down to cover the max expected extent of aggregated data in the 4
source cols C to F. If you expect source data per col to span up to say 100
rows down, then copy down by a total of 400 rows (4 cols x 100 rows = 400
rows). Col G functions to merge the contents of cols C to F into a single
col. Empty source cells will be returned as zeros.

In H5:
=IF(G5=0,"",IF(COUNTIF(G$5:G5,G5)1,"",ROW()))
Leave H1:H4 blank. This is a criteria col to ignore the zeros and flag the
unique items extracted in col G.

In I5:
=IF(ROWS($1:1)COUNT(H:H),"",INDEX(G:G,SMALL(H:H,R OWS($1:1))))

In J5:
=IF(I5="","",COUNTIF(G:G,I5))
Select H5:J5, fill down to the same extent as for col G. Hide away cols G &
H. Col I will return the list of unique items from cols C to F, while col J
returns the counts of these unique items. All results will be neatly bunched
at the top.

You could then just do a simple copy n paste special as values for cols I &
J elsewhere, and use Data Sort menu to sort by the uniques col.
--
Max
Singapore
http://savefile.com/projects/236895
xdemechanik
---
"partsman_ba" wrote:
What I would like on, say, Sheet2 would be:

1791XE 1
3231 or 3411 8
6500 1
6861 7
6862 1
7182 7

and if I add a new number to Sheet1, it shows up on my list

"partsman_ba" wrote:

Roger,

Here's a small piece of what the top looks like:

4J FILTERS

UNIT # YR/MAKE LUBE AIR FUEL
PRIMARY SECONDARY INNER
10 07 BLUEBIRD 7182 6861 3231 or 3411
11 07 BLUEBIRD 7182 6861 3231 or 3411
12 07 BLUEBIRD 7182 6861 3231 or 3411
13 07 BLUEBIRD 7182 6861 3231 or 3411
14 08 BLUEBIRD 7182 6861 3231 or 3411
15 08 BLUEBIRD 7182 6861 3231 or 3411
20 03 BLUEBIRD 7182 6500 3231 or 3411
21 06 BLUEBIRD 1791XE 6861 6862 3231 or 3411

So, not all fields are full, and headings are split - can it still work?



"Roger Govier" wrote:

Hi

Provided each of your columns has a header title in row 1, it sounds an
ideal candidate for a Pivot Table report.
Take a look at Debra Dalgleish's site and scroll down to Pivot Tables
for assistance on how to set one up.
http://www.contextures.com/tiptech.html

or Mike Alexander's site
http://www.datapigtechnologies.com/f...es/pivot1.html
--
Regards

Roger Govier


"partsman_ba" wrote in message
...
Ideally, I would like it to automatically update - if I put in a new
filter
number, it would be added to the sorted list with an occurence of 1,
or if I
add an already used filter to a new vehicle, it will add another
occurence to
the list. Am I asking too much?

"partsman_ba" wrote:

Here's what I have:

A spreadsheet sent to me with a list of filter part numbers, with
each
vehicle having it's own row and each type of filter (air, oil, etc)
having
it's own column. I would like to create from this data a sorted list
of each
different filter part number, along with a count of how many times
that part
number is on the sheet. Part numbers contain both numeric and alpha
characters in some cases. It seems like something not so hard to do,
but I've
been bashing my head for some time!