--- title: "Join" teaching: 3000 exercises: 0 questions: - "How do we use `JOIN` to combine information from multiple tables?" objectives: - "Upload a table to the Gaia server." - "Write ADQL queries involving `JOIN` operations." keypoints: - "Use `JOIN` operations to combine data from multiple tables in a databased, using some kind of identifier to match up records from one table with records from another." - "This is another example of a practice we saw in the previous notebook, moving the computation to the data." --- {% include links.md %}

Joining Tables

This is the fifth in a series of notebooks related to astronomy data.

As a continuing example, we will replicate part of the analysis in a recent paper, "Off the beaten path: Gaia reveals GD-1 stars outside of the main stream" by Adrian M. Price-Whelan and Ana Bonaca.

Picking up where we left off, the next step in the analysis is to select candidate stars based on photometry data. The following figure from the paper is a color-magnitude diagram for the stars selected based on proper motion:

In red is a stellar isochrone, showing where we expect the stars in GD-1 to fall based on the metallicity and age of their original globular cluster.

By selecting stars in the shaded area, we can further distinguish the main sequence of GD-1 from younger background stars.


Here are the steps in this notebook:

  1. We'll reload the candidate stars we identified in the previous notebook.

  2. Then we'll run a query on the Gaia server that uploads the table of candidates and uses a JOIN operation to select photometry data for the candidate stars.

  3. We'll write the results to a file for use in the next notebook.

After completing this lesson, you should be able to

  • Upload a table to the Gaia server.

  • Write ADQL queries involving JOIN operations.

Installing libraries

If you are running this notebook on Colab, you can run the following cell to install the libraries we'll use.

If you are running this notebook on your own computer, you might have to install these libraries yourself. See the instructions in the preface.

In [1]:
# If we're running on Colab, install libraries

import sys
IN_COLAB = 'google.colab' in sys.modules

    !pip install astroquery wget

Reloading the data

The following cell downloads the data from the previous notebook.

In [2]:
import os
from wget import download

filename = 'gd1_candidates.hdf5'
path = 'https://github.com/AllenDowney/AstronomicalData/raw/main/data/'

if not os.path.exists(filename):

And we can read it back.

In [3]:
import pandas as pd

candidate_df = pd.read_hdf(filename, 'candidate_df')

candidate_df is the Pandas DataFrame that contains results from the query in the previous notebook, which selects stars likely to be in GD-1 based on proper motion. It also includes position and proper motion transformed to the ICRS frame.

In [4]:
import matplotlib.pyplot as plt

x = candidate_df['phi1']
y = candidate_df['phi2']

plt.plot(x, y, 'ko', markersize=0.3, alpha=0.3)

plt.xlabel('ra (degree GD1)')
plt.ylabel('dec (degree GD1)');

This is the same figure we saw at the end of the previous notebook. GD-1 is visible against the background stars, but we will be able to see it more clearly after selecting based on photometry data.

Getting photometry data

The Gaia dataset contains some photometry data, including the variable bp_rp, which we used in the original query to select stars with BP - RP color between -0.75 and 2.

Selecting stars with bp-rp less than 2 excludes many class M dwarf stars, which are low temperature, low luminosity. A star like that at GD-1's distance would be hard to detect, so if it is detected, it it more likely to be in the foreground.

Now, to select stars with the age and metal richness we expect in GD-1, we will use g - i color and apparent g-band magnitude, which are available from the Pan-STARRS survey.

Conveniently, the Gaia server provides data from Pan-STARRS as a table in the same database we have been using, so we can access it by making ADQL queries.

In general, looking up a star from the Gaia catalog and finding the corresponding star in the Pan-STARRS catalog is not easy. This kind of cross matching is not always possible, because a star might appear in one catalog and not the other. And even when both stars are present, there might not be a clear one-to-one relationship between stars in the two catalogs.

Fortunately, smart people have worked on this problem, and the Gaia database includes cross-matching tables that suggest a best neighbor in the Pan-STARRS catalog for many stars in the Gaia catalog.

This document describes the cross matching process. Briefly, it uses a cone search to find possible matches in approximately the right position, then uses attributes like color and magnitude to choose pairs of observations most likely to be the same star.

Joining tables

So the hard part of cross-matching has been done for us. Using the results is a little tricky, but it gives us a chance to learn about one of the most important tools for working with databases: "joining" tables.

In general, a "join" is an operation where you match up records from one table with records from another table using as a "key" a piece of information that is common to both tables, usually some kind of ID code.

In this example:

  • Stars in the Gaia dataset are identified by source_id.

  • Stars in the Pan-STARRS dataset are identified by obj_id.

For each candidate star we have selected so far, we have the source_id; the goal is to find the obj_id for the same star (we hope) in the Pan-STARRS catalog.

To do that we will:

  1. Make a table that contains the source_id for each candidate star and upload the table to the Gaia server;

  2. Use the JOIN operator to look up each source_id in the gaiadr2.panstarrs1_best_neighbour table, which contains the obj_id of the best match for each star in the Gaia catalog; then

  3. Use the JOIN operator again to look up each obj_id in the panstarrs1_original_valid table, which contains the Pan-STARRS photometry data we want.

Let's start with the first step, uploading a table.

Preparing a table for uploading

For each candidate star, we want to find the corresponding row in the gaiadr2.panstarrs1_best_neighbour table.

In order to do that, we have to:

  1. Write the table in a local file as an XML VOTable, which is a format suitable for transmitting a table over a network.

  2. Write an ADQL query that refers to the uploaded table.

  3. Change the way we submit the job so it uploads the table before running the query.

The first step is not too difficult because Astropy provides a function called writeto that can write a Table in XML.

The documentation of this process is here.

First we have to convert our Pandas DataFrame to an Astropy Table.

In [5]:
from astropy.table import Table

candidate_table = Table.from_pandas(candidate_df)

To write the file, we can use Table.write with format='votable', as described here.

In [6]:
table_id = candidate_table[['source_id']]
table_id.write('candidate_df.xml', format='votable', overwrite=True)

Notice that we select a single column from the table, source_id. We could write the entire table to a file, but that would take longer to transmit over the network, and we really only need one column.

This process, taking a structure like a Table and translating it into a form that can be transmitted over a network, is called serialization.

XML is one of the most common serialization formats. One nice feature is that XML data is plain text, as opposed to binary digits, so you can read the file we just wrote:

In [7]:
!head candidate_df.xml
<?xml version="1.0" encoding="utf-8"?>
<!-- Produced with astropy.io.votable version 4.2
     http://www.astropy.org/ -->
<VOTABLE version="1.4" xmlns="http://www.ivoa.net/xml/VOTable/v1.4" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="http://www.ivoa.net/xml/VOTable/v1.4">
 <RESOURCE type="results">
   <FIELD ID="source_id" datatype="long" name="source_id"/>

XML is a general format, so different XML files contain different kinds of data. In order to read an XML file, it's not enough to know that it's XML; you also have to know the data format, which is called a schema.

In this example, the schema is VOTable; notice that one of the first tags in the file specifies the schema, and even includes the URL where you can get its definition.

So this is an example of a self-documenting format.

A drawback of XML is that it tends to be big, which is why we wrote just the source_id column rather than the whole table. The size of the file is about 750 KB, so that's not too bad.

In [8]:
!ls -lh candidate_df.xml
-rw-rw-r-- 1 downey downey 396K Dec 29 11:50 candidate_df.xml

If you are using Windows, ls might not work; in that case, try:

!dir candidate_df.xml


There's a gotcha here we want to warn you about. Why do you think we used double brackets to specify the column we wanted? What happens if you use single brackets?

Run these code snippets to find out.

table_id = candidate_table[['source_id']]
column = candidate_table['source_id']
# This one should cause an error
In [9]:
# Solution

# table_id is a Table

# column is a Column

# Column does not provide `write`, so you get an AttributeError

Uploading a table

The next step is to upload this table to the Gaia server and use it as part of a query.

Here's the documentation that explains how to run a query with an uploaded table.

In the spirit of incremental development and testing, let's start with the simplest possible query.

In [10]:
query = """SELECT *
FROM tap_upload.candidate_df

This query downloads all rows and all columns from the uploaded table. The name of the table has two parts: tap_upload specifies a table that was uploaded using TAP+ (remember that's the name of the protocol we're using to talk to the Gaia server).

And candidate_df is the name of the table, which we get to choose (unlike tap_upload, which we didn't get to choose).

Here's how we run the query:

In [11]:
from astroquery.gaia import Gaia

job = Gaia.launch_job_async(query=query, 
Created TAP+ (v1.2.1) - Connection:
	Host: gea.esac.esa.int
	Use HTTPS: True
	Port: 443
	SSL Port: 443
Created TAP+ (v1.2.1) - Connection:
	Host: geadata.esac.esa.int
	Use HTTPS: True
	Port: 443
	SSL Port: 443
INFO: Query finished. [astroquery.utils.tap.core]

upload_resource specifies the name of the file we want to upload, which is the file we just wrote.

upload_table_name is the name we assign to this table, which is the name we used in the query.

And here are the results:

In [12]:
results = job.get_results()
Table length=7346

If things go according to plan, the result should contain the same rows and columns as the uploaded table.

In [13]:
len(table_id), len(results)
(7346, 7346)
In [14]:
In [15]:

In this example, we uploaded a table and then downloaded it again, so that's not too useful.

But now that we can upload a table, we can join it with other tables on the Gaia server.

Joining with an uploaded table

Here's the first example of a query that contains a JOIN clause.

In [16]:
query1 = """SELECT *
FROM gaiadr2.panstarrs1_best_neighbour as best
JOIN tap_upload.candidate_df as candidate_df
  ON best.source_id = candidate_df.source_id

Let's break that down one clause at a time:

  • SELECT * means we will download all columns from both tables.

  • FROM gaiadr2.panstarrs1_best_neighbour as best means that we'll get the columns from the Pan-STARRS best neighbor table, which we'll refer to using the short name best.

  • JOIN tap_upload.candidate_df as candidate_df means that we'll also get columns from the uploaded table, which we'll refer to using the short name candidate_df.

  • ON best.source_id = candidate_df.source_id specifies that we will use source_id to match up the rows from the two tables.

Here's the documentation of the best neighbor table.

Let's run the query:

In [17]:
job1 = Gaia.launch_job_async(query=query1, 
INFO: Query finished. [astroquery.utils.tap.core]

And get the results.

In [18]:
results1 = job1.get_results()
Table length=3724

This table contains all of the columns from the best neighbor table, plus the single column from the uploaded table.

In [19]:

Because one of the column names appears in both tables, the second instance of source_id has been appended with the suffix _2.

The length of results1 is about 3000, which means we were not able to find matches for all stars in the list of candidates.

In [20]:

To get more information about the matching process, we can inspect best_neighbour_multiplicity, which indicates for each star in Gaia how many stars in Pan-STARRS are equally likely matches.

In [21]:
<MaskedColumn name='best_neighbour_multiplicity' dtype='int16' description='Number of neighbours with same probability as best neighbour' length=3724>

It looks like most of the values are 1, which is good; that means that for each candidate star we have identified exactly one source in Pan-STARRS that is likely to be the same star.

To check whether there are any values other than 1, we can convert this column to a Pandas Series and use describe, which we saw in in Lesson 3.

In [22]:
import pandas as pd

multiplicity = pd.Series(results1['best_neighbour_multiplicity'])
count    3724.0
mean        1.0
std         0.0
min         1.0
25%         1.0
50%         1.0
75%         1.0
max         1.0
dtype: float64

In fact, 1 is the only value in the Series, so every candidate star has a single best match.

Similarly, number_of_mates indicates the number of other stars in Gaia that match with the same star in Pan-STARRS.

In [23]:
mates = pd.Series(results1['number_of_mates'])
count    3724.0
mean        0.0
std         0.0
min         0.0
25%         0.0
50%         0.0
75%         0.0
max         0.0
dtype: float64

All values in this column are 0, which means that for each match we found in Pan-STARRS, there are no other stars in Gaia that also match.

Detail: The table also contains number_of_neighbors which is the number of stars in Pan-STARRS that match in terms of position, before using other criteria to choose the most likely match.

Getting the photometry data

The most important column in results1 is original_ext_source_id which is the obj_id we will use to look up the likely matches in Pan-STARRS to get photometry data.

The process is similar to what we just did to look up the matches. We will:

  1. Make a table that contains source_id and original_ext_source_id.

  2. Write the table to an XML VOTable file.

  3. Write a query that joins the uploaded table with gaiadr2.panstarrs1_original_valid and selects the photometry data we want.

  4. Run the query using the uploaded table.

Since we've done everything here before, we'll do these steps as an exercise.


Select source_id and original_ext_source_id from results1 and write the resulting table as a file named external.xml.

In [24]:
# Solution

table_ext = results1[['source_id', 'original_ext_source_id']]
table_ext.write('external.xml', format='votable', overwrite=True)

Use !head to confirm that the file exists and contains an XML VOTable.

In [25]:
!head external.xml
<?xml version="1.0" encoding="utf-8"?>
<!-- Produced with astropy.io.votable version 4.2
     http://www.astropy.org/ -->
<VOTABLE version="1.4" xmlns="http://www.ivoa.net/xml/VOTable/v1.4" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="http://www.ivoa.net/xml/VOTable/v1.4">
 <RESOURCE type="results">
   <FIELD ID="source_id" datatype="long" name="source_id" ucd="meta.id;meta.main">
     Unique Gaia source identifier


Read the documentation of the Pan-STARRS table and make note of obj_id, which contains the object IDs we'll use to find the rows we want.

Write a query that uses each value of original_ext_source_id from the uploaded table to find a row in gaiadr2.panstarrs1_original_valid with the same value in obj_id, and select all columns from both tables.

Suggestion: Develop and test your query incrementally. For example:

  1. Write a query that downloads all columns from the uploaded table. Test to make sure we can read the uploaded table.

  2. Write a query that downloads the first 10 rows from gaiadr2.panstarrs1_original_valid. Test to make sure we can access Pan-STARRS data.

  3. Write a query that joins the two tables and selects all columns. Test that the join works as expected.

As a bonus exercise, write a query that joins the two tables and selects just the columns we need:

  • source_id from the uploaded table

  • g_mean_psf_mag from gaiadr2.panstarrs1_original_valid

  • i_mean_psf_mag from gaiadr2.panstarrs1_original_valid

Hint: When you select a column from a join, you have to specify which table the column is in.

In [26]:
# Solution

# First test

query2 = """SELECT *
FROM tap_upload.external as external

# Second test

query2 = """SELECT TOP 10
FROM gaiadr2.panstarrs1_original_valid

# Third test

query2 = """SELECT *
FROM gaiadr2.panstarrs1_original_valid as ps
JOIN tap_upload.external as external
  ON ps.obj_id = external.original_ext_source_id

# Complete query

query2 = """SELECT
external.source_id, ps.g_mean_psf_mag, ps.i_mean_psf_mag
FROM gaiadr2.panstarrs1_original_valid as ps
JOIN tap_upload.external as external
  ON ps.obj_id = external.original_ext_source_id

Here's how we launch the job and get the results.

In [27]:
job2 = Gaia.launch_job_async(query=query2, 
INFO: Query finished. [astroquery.utils.tap.core]
In [28]:
results2 = job2.get_results()
Table length=3724


Optional Challenge: Do both joins in one query.

There's an example here you could start with.

In [30]:
# Solution

query3 = """SELECT
candidate_df.source_id, ps.g_mean_psf_mag, ps.i_mean_psf_mag
FROM tap_upload.candidate_df as candidate_df
JOIN gaiadr2.panstarrs1_best_neighbour as best
  ON best.source_id = candidate_df.source_id
JOIN gaiadr2.panstarrs1_original_valid as ps
  ON ps.obj_id = best.original_ext_source_id

# job3 = Gaia.launch_job_async(query=query3, 
#                       upload_resource='candidate_df.xml', 
#                       upload_table_name='candidate_df')

# results3 = job3.get_results()
# results3

Write the data

Since we have the data in an Astropy Table, let's store it in a FITS file.

In [31]:
filename = 'gd1_photo.fits'
results2.write(filename, overwrite=True)

We can check that the file exists, and see how big it is.

In [32]:
!ls -lh gd1_photo.fits
-rw-rw-r-- 1 downey downey 96K Dec 29 11:51 gd1_photo.fits

At around 175 KB, it is smaller than some of the other files we've been working with.

If you are using Windows, ls might not work; in that case, try:

!dir gd1_photo.fits


In this notebook, we used database JOIN operations to select photometry data for the stars we've identified as candidates to be in GD-1.

In the next notebook, we'll use this data for a second round of selection, identifying stars that have photometry data consistent with GD-1.

Best practice

  • Use JOIN operations to combine data from multiple tables in a databased, using some kind of identifier to match up records from one table with records from another.

  • This is another example of a practice we saw in the previous notebook, moving the computation to the data.

In [ ]: