Database Notes
The simplest approach would be to place all pertinent information about each frame in a single "row" of the database. A superior strategy seems to be the creation of a cluster of cross-referencing databases.
To start, we make a small database containing the site specific information. Each site is uniquely identified by a string code or integer ID. However, there may be multiple records with the same ID but different dates, as certain information may change with time. The last field of each record will contain a list of instrument IDs at each site.
| Sites | siteID | date | siteName | siteCode | latitude | longitude | PACE latitude | instrumentIDs |
|---|---|---|---|---|---|---|---|
| 000 | 2001-09-12 | University of Calgary | UofC | 56.38 | 265.35 | 66.95 | null |
| 001 | 2001-09-12 | Gillam | GILL | 56.38 | 265.35 | 66.95 | 003 004 009 |
| 001 | 2002-09-01 | Gillam | GILL | 56.38 | 265.35 | 66.97 | 004 009 |
| 002 | 2001-09-12 | Rankin | RANK | 62.82 | 267.89 | 73.24 | 012 |
The second database contains all instrument specific details. As with sites, each instrument has a unique integer ID and a date stamp to allow for modifications and upgrades. The siteID corresponds to the unique identifier from the site database. There is also a list of filters.
| Instruments | instrumentID | date | name | type | siteID | ASIfilterIDs |
|---|---|---|---|---|---|
| 000 | 2001-09-12 | Darkroom | null | 000 | null |
| 003 | 2001-09-12 | Wilbur | ASI | 001 | 003 004 012 015 019 |
| 004 | 2001-09-12 | Polaris | ASI | 001 | 021 022 024 025 029 |
| 012 | 2001-09-12 | Aqsaniq | ASI | 001 | 041 045 046 047 |
Filter information is located in a third database. The basic idea is the same as for sites and instruments. Each filter will typically be assocated with an instrument via the corresponding ID number. Maybe also add an array containing the entire transmission curve versus wavelength?
| Filters | filterID | date | center Wavelength | band Width | peak Transmission | instrumentID |
|---|---|---|---|---|---|
| 003 | 2001-12-23 | 630.0 | 1.8 | 0.89 | 04 |
| 004 | 2001-12-23 | 557.7 | 2.0 | 0.45 | 04 |
These three small databases should be relatively easy to maintain. They could also be printed out in tabular form for easy access. More importantly, they simplify the records in an image frame database. Each frame record will first be tagged with a unique identifier. This differs from the other databases in that there will only be one frame for each frameID. Date and time of exposure are also included. Next, instrument and filter IDs refer to the corresponding entries in those databases. This is followed by additional frame specific information such as integration time. We should endevour to minimize the size of each frame record, so that the database will remain useful even with several million records. If database size does become a problem then it might be worth splitting it up into a separate database for each site.
| Frames | frameID | date | time | siteID | instrumentID | filterID | exposureTime |
|---|---|---|---|---|---|---|
| 000001 | 2001-12-23 | 06:12:32 | 01 | 02 | 3 | 1200 ms |
| 000002 | 2001-12-23 | 06:12:36 | 01 | 02 | 4 | 100 ms |
A list of stuff that might be overhead
| Satellites | satelliteID | name | launchDate | lastDate | otherstuff |
|---|---|---|---|---|
| 103 | FAST | 1999-09-12 | 2004-01-01 | ??? |
| 103 | FAST | 1999-09-12 | 2004-01-01 | ??? |
| 103 | FAST | 1999-09-12 | 2004-01-01 | ??? |
could be used to generate (from SSCWeb) a table of overflights
| Overflights | date | time | satelliteID | siteID | distance |
|---|---|---|---|---|
| 1999-09-12 | 11:32:35 | 103 | 01 | 450 km |
| 1999-09-12 | 11:34:12 | 103 | 02 | 121 km |
