I (dstn) decided to get some good stats about how hard it is to solve GALEX.

It turns out that GALEX releases the results of running SExtractor on their tile data, so I gathered these.

Results

Total number of fields: 7077.

NUV:

Run	Fields Solved	Fields Solved (%)	Fields Unsolved	Fields Unsolved (%)
109-1	7021	99.21	56	0.79
109-2	7064	99.82	13	0.18
109-3	7068	99.87	9	0.13
109-5	7069	99.89	8	0.11

---

Below here lie dragons of purely archival interest.

---

Data Gathering

I collected the data as follows:

• Starting at http://galex.stsci.edu/GR2/Default.aspx?page=tilelist&survey=ais , I clicked the "Show All Tiles" button.
• Clicked the "Tilenum" column header to sort the tiles by tile number.
• Went through all 71 pages and did "Save As...", naming the files galex01.html through galex71.html; these are in raid2/GALEX_FIELDS/html.
• Ran the following awk script (awk -f get-galex-url.awk html/galex*.html > nuv/urls):

  BEGIN{RS="[[:space:]=]"}
  /xd-int_2color_medium\.jpg">/{sub("/qa",""); sub("xd-int_2color_medium.jpg", "nd-cat.fits.gz"); sub("\"", ""); sub("\">", ""); print}
  

• Ran awk '{sub("-nd-", "-fd-"); print;}' nuv/urls > fuv/urls to create the FUV URL list.
• Ran xargs wget < nuv/urls and xargs wget < fuv/urls to download the files.
• Ran the following to get out the sorted filename list: awk 'BEGIN{FS="/"} {print $NF}' < nuv/urls > nuv/files and the same for FUV.
• Ran the following shell script (cd nuv; reduce < files; cd ../fuv; reduce < files)to convert the files to xylists:

  #! /bin/bash
  n=0;
  for ((;;)); do
          read x
          echo $x
          N=`printf %04i $n`
          gunzip -cd $x > field.fits
          tabsort -i field.fits -o sorted.fits -c MAG_AUTO
          fitscopy sorted.fits"[#row<=300][col X=X_IMAGE;Y=Y_IMAGE;MAG_AUTO]" nuv$N.fits
          rm field.fits
          rm sorted.fits
          n=$(($n+1))
  done
  

• Grabbed a basic FITS header:

  dd if=~/raid3/SDSS_FIELDS_UGRIZ/sdssfield_i01.xy3.fits of=hdr bs=1 count=2880
  

• Ran a script to pull out the FITS tables from the files:

  #! /bin/bash
  mkdir tmp
  for ((x=0; x<7077; x++)); do
          echo $x
          IN=`printf fuv%04i.fits $x`
          OUT=`printf tmp/%04i $x`
          TILE=`modhead $IN TILENUM | awk '{print $3}'`
          OBJ=`modhead $IN OBJECT | awk '{gsub("'\''", ""); print $3}'`
          ID=`echo FUV-$OBJ`
          echo $ID
          fitsgetext -i "$IN" -o "$OUT" -e 1
          modhead "$OUT" SURVEY GALEX
          modhead "$OUT" FIELDID "$ID"
          modhead "$OUT" TILENUM "$TILE"
  done
  

• Catted them all together:

  cat hdr tmp/* > fuv.xy.fits
  

• Pulled out the RA,DEC values (reduce-rd):

  #! /bin/bash
  n=0;
  for ((;;)); do
          read x
          echo $x
          x="../nuv-raw/$x"
          N=`printf %04i $n`
          gunzip -cd $x > field.fits
          tabsort -i field.fits -o sorted.fits -c MAG_AUTO
          fitscopy sorted.fits"[#row<=300][col RA=ALPHA_J2000;DEC=DELTA_J2000]" nuv$N.rd.fits
          rm field.fits
          rm sorted.fits
          n=$(($n+1))
  done
  

• Grabbed the first extensions:

  #! /bin/bash
  mkdir tmp
  for ((x=0; x<7077; x++)); do
          echo $x
          IN=`printf nuv%04i.rd.fits $x`
          OUT=`printf tmp/%04i $x`
          TILE=`modhead $IN TILENUM | awk '{print $3}'`
          OBJ=`modhead $IN OBJECT | awk '{gsub("'\''", ""); print $3}'`
          ID=`echo NUV-$OBJ`
          echo $ID
          fitsgetext -i "$IN" -o "$OUT" -e 1
          modhead "$OUT" SURVEY GALEX
          modhead "$OUT" FIELDID "$ID"
          modhead "$OUT" TILENUM "$TILE"
  done
  

• Catted them all together:

  cat hdr tmp/* > nuv.rd.fits