I used wc -l file.csv
to count the records in a tab-delimited file. Shows 127992
.
Connect to file in Tableau and it shows 63k records and some change. OK, so I check for blank IDs in the source file:
awk -F'\t' '!length($1)' file.csv
awk -F'\t' '!length($1){print $0}'
awk -F'\t' '$1 == ""{print $0}'
In no case did it expose a null or blank ID, which would be very low probability as the database this is pulled from doesn't allow this to be null. Trying to figure out, in what use case scenario would you want your software to discard half the data?
In Tableau, I create a calculated field, wondering why this would be necessary, but OK:
COUNT([Itemnumber])
Keep in mind, I double-checked to make sure there were no NULL values and I get: 63,756
I'm used to Unix and new to Tableau so what am I missing?
Sample rows from file:
itemnumber title author itemcallnumber barcode lib datelastseen datelastborrowed amountoutstanding
7666 Indian legends of American scenes / Gridley, Marion E. J 398.2 GRIDLEY 39030000964991 Lost 2001-06-04 2001-06-04
12927 The legend of the Indian paintbrush / DePaola, Tomie, J 398.2089 DE PAOL 39030002344226 Lost 2002-04-03 2002-04-03
16943 The Tainos : Jacobs, Francine. J 972.901 JACOBS 39030001423195 Lost 2003-06-04 2003-06-04
124564 Columbia River, the Astoria odyssey : Penttila, Bryan. 979.7002 PENTTIL 39030008124408 Lost 2005-09-12 2005-09-12