pjoin(1)(TDH) pjoin(1)(TDH)
NAMEpjoin(1) - relational join on two files
SYNOPSISpjoin [options] file1 keyfields1 file2 keyfields2
DESCRIPTIONpjoin performs a relational join on two sets of whitespace-delimited
tabular data records. file1 and file2 are the input files. One of
them may be - to indicate that data are to be read from standard input.
The result is written to standard output and uses the same field delim‐
itation as the input files. Comment lines (beginning with //) and
blank lines are skipped on input and do not appear in output.
keyfields1 specifies one or more fields in file1 to consider when per‐
forming the join; it may be a single field specifier or a list of field
specifiers delimited by commas. Likewise keyfields2 for file2. Fields
may be specified by number, e.g. 2 specifies the second field. If a
file has a field name header, field names may be used, e.g. id (how‐
ever the -h1 and/or -h2 options must be used so that pjoin knows to
expect a field name header in file1 or file2 respectively).
OPTIONS-l
Do a left join. Prevents records from being omitted from the
left side (file1) as the result of the join. Missing records
will be filled with placeholder characters.
-r
Do a right join. Prevents records from being omitted from the
right side (file2) as the result of the join. Missing records
will be filled with placeholder characters.
Note: -l and -r may both be used to produce a loss-less join.
-i
Make comparisons case-insensitive. Normally they are case-sen‐
sitive.
-dup1
Allow multiple instances in file1. The matching record from
file2 will be replicated for each instance.
-dup2
Allow multiple instances in file2. The matching record from
file1 will be replicated for each instance.
-q
Quick option. Do not sort input; assume inputs are already in
sort order. Normally the inputs are piped through an appropri‐
ate sort(1) command to sort on the key fields. pjoin will not
give correct results if inputs are unsorted.
-rml
Remove the lefthand portion of the result, leaving only the
records from file2.
-rmr
Remove the righthand portion of the result, leaving only the
records from file1.
-H
Both file1 and file2 have field name headers, and a field name
header will also be written to the output. Equivalent to -h1
-h2 -ho.
-h1
Indicates that file1 has a field name header. This allows
fields in file1 to be specified by name.
-h2
Indicates that file2 has a field name header. This allows
fields in file2 to be specified by name.
-ho
Field name header will be written as the first line of output.
At least one of the input files must have a field name header.
If one of the files did not have a field name header, place‐
holder fill characters will be written for that portion of the
field name header.
-t
Indicates that input and output are tab delimited. Normally the
join result uses a space between the left side and right side;
with -t a tab is used instead.
-fc
Set the placeholder fill character to c. Normally it is -.
NOTES
Using -l anf -r together results in a "loss-less" join.
-dup1 and -dup2 cannot be used together.
-rmr and -rml cannot be used together.
EXAMPLE
Suppose file1 looks like this:
001 A red
001 B red
002 A blue
003 C yellow
and file2 looks like this:
001 A Jean
002 A Jan
We could perform an ordinary join by issuing the command: pjoin file1
1,2 file2 1,2
001 A red 001 A Jean
002 A blue 002 A Jan
Or we could perform a left join by issuing this command: pjoin-l file1
1,2 file2 1,2
001 A red 001 A Jean
001 B red --- - ----
002 A blue 002 A Jan
003 C yellow --- - ---
AUTHOR
Steve Grubb, with portions developed by Sandra Reynolds and Marv New‐
house.
22-SEP-2003 TDH scg@jax.org pjoin(1)(TDH)
