...below a redshift of 0.1:
nsa.z < 0.1
...on plates 7815 and 8485:
cube.plate == 7815 or cube.plate == 8485
...with a IFU size of 127:
ifu.name = 127*
...with g-r color < 0.2:
g_r < 0.2
...with Halpha flux > 25 in more than 20% of their good spaxels:
npergood(emline_gflux_ha_6564 > 25) >= 20
...that have an Halpha flux > 25:
emline_gflux_ha_6564 > 25
...that have Halpha EW > 6 in galaxies between stellar mass 9.5-11 with sersic index < 2:
nsa.sersic_logmass >= 9.5 and nsa.sersic_logmass < 11 and nsa.sersic_n < 2 and emline_sew_ha_6564 > 6
...that have NII/Halpha ratio > 0.1:
nii_to_ha > 0.1
emline_gflux_ha_6564 > 25
npergood(spaxelprop.emline_gflux_ha_6564 > 5) >= 20
nsa.sersic_logmass >= 9.5 and nsa.sersic_logmass < 11 and nsa.sersic_n < 2 and emline_sew_ha_6564 > 6
nsa.z < 0.1 and haflux > 25
Query timing exercise post postgres 9.3 config optimization. From a laptop in Baltimore to Utah with explicit queries in remote mode
Query | 1st Run (s) | 2nd Run (s) | Best of 5 (1 loop) | Best of 5 (10 loops) |
---|---|---|---|---|
Q1 | 0.3012 | 0.2375 | 682 ms per loop | 710 ms per loop |
Q2 | fail | fail | ||
Q3 | 39.493 | 10.087 | 7 s per loop | 7.19 s per loop |
Q4 | 1.1662 | 0.3331 | 783 ms per loop | 810 ms per loop |
q = Query(search_filter=p, mode='remote') # Best of 5 (1 loop) %timeit -n 1 -r 5 r = q.run() # Best of 5 (10 loops) %timeit -n 10 -r 5 r = q.run()
Query timing from pasting the raw SQL into psql at Utah, pre postgres 9.6 config optimization.
Query | 1st Run (s) | 2nd Run (s) | Best of 5 (1 loop) |
---|---|---|---|
Q1 | 16.994 | 0.9450 | 679 ms per loop |
Q2 | 285.67 | 82.263 | 30.9 s per loop |
Q3 | 234.72 | 35.032 | 33.7 s per loop |
Q4 | 0.7135 | 0.6954 | 716 ms per loop |
Query timing from pasting the raw SQL into psql at Utah, post postgres 9.6 config optimization + parallel queries
Query | 1st Run (s) | 2nd Run (s) | Best of 5 (1 loop) |
---|---|---|---|
Q1 | 1.0 | 0.517 | 514 ms per loop |
Q2 | 31.123 | 11.120 | 8.56 s per loop |
Q3 | 81.319 | 25.612 | 25.4 s per loop |
Q4 | 0.588 | 0.5668 | 566 ms per loop |