I need to extract the row of data where the absolute difference of y(block_1)-pt is less than a certain limit. If more than one row satisfies this condition, I need to retrieve the row with the smallest difference.
$Data <<EOD
# block_1 # block_2
x y x y
1 1979.2 2 2027
1 1598.6 2 1591.4
1 1430.8 2 1401.6
1 1544.7 2 1923.7
1 1953.3 2 1677
1 1288.8 2 1311.1
1 1741.3 2 1689.3
1 1791.1 2 2015.8
EOD
set key autotitle columnhead
pt = 1953.1
limit = 0.5
absDiff(col,x) = abs( column(col) - x ) < limit
set table $Temp
plot $Data u ( absDiff( 2, pt ) ):( $4 ) w table
unset table
print $Temp
The result is:
0 2027
0 1591.4
0 1401.6
0 1923.7
1 1677
0 1311.1
0 1689.3
0 2015.8
How can I skip all rows that contain a 0 in the first column?
If you are using gnuplot>5.4.0 you can use the experimental feature "conditional writing to a table" (check help with table). If you are bound to versions <5.4.0, there are other solutions.
If you change your plot with table command to
plot $Data u ( absDiff( 2, pt ) ):4 w table if absDiff( 2, pt )
the result will be:
1 1677
Addition: finding the minimum difference without giving a limit and using stats.
You better skip the line x y x y because it is no valid data line and it will shift the row index by one. Row indices are zero-based.
Script: (requires gnuplot>=5.0, because of use of datablocks, stats works also with earlier 4.x versions)
### find minimum difference and corresponding row
reset session
$Data <<EOD
# block_1 # block_2
x y x y
1 1979.2 2 2027
1 1598.6 2 1591.4
1 1430.8 2 1401.6
1 1544.7 2 1923.7
1 1953.3 2 1677
1 1288.8 2 1311.1
1 1741.3 2 1689.3
1 1791.1 2 2015.8
EOD
pt = 1953.1
stats $Data u 4:(abs($2-pt)) skip 1 nooutput
print sprintf("Minimum difference is: %g",STATS_min_y)
print sprintf("Located in row: %g",STATS_index_min_y)
print sprintf("Value in column 4 is: %g",STATS_pos_min_y)
### end of script
Result:
Minimum difference is: 0.2
Located in row: 4
Value in column 4 is: 1677