Any sed option to stop matching when one line number is reached?

Question

I have a table where only even rows are important (bold ones),

year	month	day	time	lat	lon	value
2020	1	1	00:00:00	41.97	29.68	272.4456
2020	1	1	00:00:00	47.97	21.68	-32767
2020	1	1	01:00:00	41.97	29.68	272.3355
2020	1	1	01:00:00	47.97	21.68	-32767
2020	1	1	02:00:00	41.97	29.68	272.1232
2020	1	1	02:00:00	47.97	21.68	-32767
2020	1	1	03:00:00	41.97	29.68	271.8801
2020	1	1	03:00:00	47.97	21.68	-32767

but, at some point, this changes to odd rows. This is the exact moment where it changes (you can see two consecutive "-32767"):

year	month	day	time	lat	lon	value
2023	9	30	23:00:00	41.97	29.68	289.2723
2023	9	30	23:00:00	41.97	29.68	-32767
2023	10	1	00:00:00	41.97	29.68	-32767
2023	10	1	00:00:00	41.97	29.68	288.9512
2023	10	1	01:00:00	41.97	29.68	-32767
2023	10	1	01:00:00	41.97	29.68	288.7689
2023	10	1	02:00:00	41.97	29.68	-32767

I found this out after extracting them with sed '1~2d' site8_slt.nc.tsv. Does anyone know how to sed until a specific line number? This way I would sed two times, first before the change and then after the change.

Thanks in advance!

Context: I am downloading data from cds.climate.copernicus.eu in netCDF format and then converting the data to csv with cdo (Climate Data Operators), and, I dont know why, I get 2 values per row where each value is always the fill value -32767.

I have tried: I know that I can point to specific lines with this $ sed '1s/a/b/' file but this only changes one line, I want to to sed all lines before one in a way, and then sed in another way after one specific line number.

Answer 1

To answer your specific question:

$ seq 5 | sed -n '1,3p'
1
2
3

$ seq 5 | sed -n '4,$p'
4
5

but it sounds like either of these might be more useful for you, using any sed or any awk:

sed '/-32767$/d' file
awk '$NF != -32767' file

For example:

$ cat file
year    month   day     time    lat     lon     value
2020    1       1       00:00:00        41.97   29.68   272.4456
2020    1       1       00:00:00        47.97   21.68   -32767
2020    1       1       01:00:00        41.97   29.68   272.3355
2020    1       1       01:00:00        47.97   21.68   -32767
2020    1       1       02:00:00        41.97   29.68   272.1232
2020    1       1       02:00:00        47.97   21.68   -32767
2020    1       1       03:00:00        41.97   29.68   271.8801
2020    1       1       03:00:00        47.97   21.68   -32767
2023    9       30      23:00:00        41.97   29.68   289.2723
2023    9       30      23:00:00        41.97   29.68   -32767
2023    10      1       00:00:00        41.97   29.68   -32767
2023    10      1       00:00:00        41.97   29.68   288.9512
2023    10      1       01:00:00        41.97   29.68   -32767
2023    10      1       01:00:00        41.97   29.68   288.7689
2023    10      1       02:00:00        41.97   29.68   -32767

$ sed '/-32767$/d' file
year    month   day     time    lat     lon     value
2020    1       1       00:00:00        41.97   29.68   272.4456
2020    1       1       01:00:00        41.97   29.68   272.3355
2020    1       1       02:00:00        41.97   29.68   272.1232
2020    1       1       03:00:00        41.97   29.68   271.8801
2023    9       30      23:00:00        41.97   29.68   289.2723
2023    10      1       00:00:00        41.97   29.68   288.9512
2023    10      1       01:00:00        41.97   29.68   288.7689

$ awk '$NF != -32767' file
year    month   day     time    lat     lon     value
2020    1       1       00:00:00        41.97   29.68   272.4456
2020    1       1       01:00:00        41.97   29.68   272.3355
2020    1       1       02:00:00        41.97   29.68   272.1232
2020    1       1       03:00:00        41.97   29.68   271.8801
2023    9       30      23:00:00        41.97   29.68   289.2723
2023    10      1       00:00:00        41.97   29.68   288.9512
2023    10      1       01:00:00        41.97   29.68   288.7689

awk will almost certainly be much more useful to you than sed in doing whatever else you want to do with this data.