Tables

Synapse Tables enable storage of tabular data in Synapse in a form that can be queried using a SQL-like query language.

A table has a Schema and holds a set of rows conforming to that schema.

A Schema defines a series of Column of the following types: STRING, DOUBLE, INTEGER, BOOLEAN, DATE, ENTITYID, FILEHANDLEID, LINK, LARGETEXT, USERID

Preliminaries:

Let’s say we have some data stored in an R data.frame:

genes <- data.frame(
  Name = c("foo", "arg", "zap", "bah", "bnk", "xyz"),
  Chromosome = c(1, 2, 2, 1, 1, 1),
  Start = c(12345, 20001, 30033, 40444, 51234, 61234),
  End = c(126000, 20200, 30999, 41444, 54567, 68686),
  Strand = c("+", "+", "-", "-", "+", "+"),
  TranscriptionFactor = c(F, F, F, F, T, F),
  Time = as.POSIXlt(c("2017-02-14 11:23:11.024", "1970-01-01 00:00:00.000", "2018-10-01 00:00:00.000", "2020-11-03 04:59:59.999", "2011-12-16 06:23:11.139", "1999-03-18 21:03:33.044"), tz = "UTC", format = "%Y-%m-%d %H:%M:%OS"))

To create a Table with name “My Favorite Genes” with project as parent:

synBuildTable creates a Table Schema based on the data and returns a Table object that can be stored in Synapse using synStore(). To create a custom Table Schema, defines the columns of the table:

cols <- list(
    Column(name = "Name", columnType = "STRING", maximumSize = 20),
    Column(name = "Chromosome", columnType = "STRING", maximumSize = 20),
    Column(name = "Start", columnType = "INTEGER"),
    Column(name = "End", columnType = "INTEGER"),
    Column(name = "Strand", columnType = "STRING", enumValues = list("+", "-"), maximumSize = 1),
    Column(name = "TranscriptionFactor", columnType = "BOOLEAN"),
    Column(name = "Time", columnType = "DATE"))

schema <- Schema(name = "My Favorite Genes", columns = cols, parent = project)

table <- Table(schema, genes)

Let’s store that in Synapse:

The Table() function takes two arguments, a schema object, or a Table ID and data in some form, which can be:

  • a path to a CSV file
  • a data frame
  • a RowSet object
  • a list of lists where each of the inner lists is a row

We now have a table populated with data. Let’s try to query:

synTableQuery() downloads the data and saves it to a csv file at location:

To load the data into an R data.frame:

Changing Data

Once the schema is settled, changes come in two flavors: appending new rows and updating existing ones.

Changing data in a table requires row IDs and version numbers for each row to be modified (called ROW_ID and ROW_VERSION). We get those by querying before updating. Minimizing change sets to contain only rows that actually change will make processing faster. Appending new rows can be accomplished by leaving values for ROW_ID and ROW_VERSION blank.

Appending new rows is fairly straightforward. To continue the previous example, we might add some new genes:

For example, let’s update the names of some of our favorite genes:

Let’s save that:

Now, query the table again to see your changes:

One other piece of information required for making changes to tables is the etag, which is used by the Synapse server to prevent concurrent users from making conflicting changes through a technique called optimistic concurrency. This comes as a result of running synTableQuery. In the example above, you could see the etag by running results$etag - but you should never have to use it directly. In case of a conflict, your update may be rejected. You then have to do another query and try your update again.

Changing Table Structure

Adding columns can be done using the methods Schema$addColumn() or addColumns() on the Schema object:

schema <- synGet(tableId)
newColumn <- synStore(Column(name = "Note", columnType = "STRING", maximumSize = 20))
schema$addColumn(newColumn)
## NULL
schema <- synStore(schema)

In the example above, we do know what newColumn is. While working with other columns, you can retrieve all columns a table has by using synGetTableColumns():

You can then explore the list of columns to find the column you want to modify.

Renaming or otherwise modifying a column involves removing the column and adding a new column:

Now we can set the values for the new column:

Updating Column Type

Column “Notes” has type STRING with “maximumSize” set to 20. We cannot add a new row with “Notes” as “a very looooooooong note” since it has more than 20 characters. Let“s change the ColumnType to”STRING" with “maximumSize” set to 100:

# getting the existing table metadata and data
originalSchema <- synGet(tableId)
oldQueryResults <- synTableQuery(sprintf("SELECT * FROM %s", tableId))
## 
 [####################]100.00%   1/1   Done...    
Downloading  [####################]100.00%   628.0bytes/628.0bytes (553.9kB/s) Job-803672087057634028769726.csv Done...
oldData <- as.data.frame(oldQueryResults)

# remove the column
originalSchema$removeColumn(notesColumn)
## NULL
newSchema <- synStore(originalSchema)

# create a new Column
newCol <- Column(name = "Notes", columnType = "STRING", maximumSize = 100)

# add the new column to the new table
newSchema$addColumn(newCol)
## NULL
newSchema <- synStore(newSchema)

# copy the data over to the new column
newQueryResults <- synTableQuery(sprintf("SELECT * FROM %s", newSchema$properties$id))
## 
Waiting for the table index to become available... [--------------------]0.00%   0/100       
Waiting for the table index to become available... [--------------------]0.00%   0/100       
Create CSV FileHandle [####################]100.00%   16/16   Done...    
Create CSV FileHandle [####################]100.00%   16/16   Done...    
Downloading  [####################]100.00%   603.0bytes/603.0bytes (581.8kB/s) Job-80368340298473666293043.csv Done...
newData <- as.data.frame(newQueryResults)
newData["Notes"] <- oldData["Notes"]

# save the change
synStore(Table(tableId, newData))
## 
Uploading [--------------------]0.00%   0.0bytes/612.0bytes  file158292dd2c12d     
Uploading [####################]100.00%   612.0bytes/612.0bytes (807.7bytes/s) file158292dd2c12d Done...    
 [####################]100.00%   1/1   Done...
## <synapseclient.table.CsvFileTable object at 0x11a27a198>

# add the new data
moreGenes <- data.frame(
    Name = c("not_sure"),
    Chromosome = c(2),
    Start = c(12345),
    End = c(126000),
    Strand = c("+"),
    TranscriptionFactor = c(F),
    Time = as.POSIXlt("2014-07-03 20:12:44.000", tz = "UTC", format = "%Y-%m-%d %H:%M:%OS"),
    Notes = c("a very looooooooong note"))
synStore(Table(tableId, moreGenes))
## 
Uploading [--------------------]0.00%   0.0bytes/165.0bytes  file15829678888db     
Uploading [####################]100.00%   165.0bytes/165.0bytes (241.9bytes/s) file15829678888db Done...    
 [####################]100.00%   1/1   Done...
## <synapseclient.table.CsvFileTable object at 0x1127a7c50>

To access a column that you do not have a reference to, please see:

Notes on Dates and Times

In Synapse tables, the DATE type is stored as a timestamp integer, equivalent to the number of milliseconds that have passed since 1970-01-01 00:00:00 UTC. R has built in POSIXt types that are similar, though they are numerics that store time in seconds. When values of type POSIXt are uploaded to Synapse tables in synapser, values are automatically converted to millisecond timestamps. Conversely, values in Synapse table columns of type DATE are automatically converted to POSIXlt by synapser.

When adding or changing dates, they may be POSIXt times.

Dates may also be submitted in timestamp milliseconds:

Note that using POSIXlt is strongly preferred over POSIXct, because POSIXct does not store values with enough precision to reliably recover milliseconds. For more information, see the R documentation:

Table Attached Files

Synapse tables support a special column type called ‘File’ which contain a file handle, an identifier of a file stored in Synapse. Here’s an example of how to upload files into Synapse, associate them with a table and read them back later:

newCols <- list(
    Column(name = "artist", columnType = "STRING", maximumSize = 50),
    Column(name = "album", columnType = "STRING", maximumSize = 50),
    Column(name = "year", columnType = "INTEGER"),
    Column(name = "catalog", columnType = "STRING", maximumSize = 50),
    Column(name = "covers", columnType = "FILEHANDLEID"))
newSchema <- synStore(Schema(name = "Jazz Albums", columns = newCols, parent = project))

newData <- data.frame(
  artist = c("John Coltrane", "Sonny Rollins", "Sonny Rollins", "Kenny Burrel"),
  album = c("Blue Train", "Vol. 2", "Newk's Time", "Kenny Burrel"),
  year = c(1957, 1957, 1958, 1956),
  catalog = c("BLP 1577", "BLP 1558", "BLP 4001", "BLP 1543")
)

# writing some temp files to upload or pointing to existing files in your system

files <- c("coltraneBlueTrain.jpg", "rollinsBN1558.jpg", "rollinsBN4001.jpg", "burrellWarholBN1543.jpg")

# upload to filehandle service
files <- lapply(files, function (f) {
  cat(f, file = f)
  synUploadSynapseManagedFileHandle(f)
  })
## 
Uploading [####################]100.00%   21.0bytes/21.0bytes  coltraneBlueTrain.jpg Done...    
Uploading [####################]100.00%   17.0bytes/17.0bytes  rollinsBN1558.jpg Done...    
Uploading [####################]100.00%   17.0bytes/17.0bytes  rollinsBN4001.jpg Done...    
Uploading [####################]100.00%   23.0bytes/23.0bytes  burrellWarholBN1543.jpg Done...

# get the filehandle ids
fileHandleIds <- lapply(files, function(f) f$id)
newData["covers"] <- fileHandleIds
## Warning in `[<-.data.frame`(`*tmp*`, "covers", value = list("609037",
## "609040", : provided 4 variables to replace 1 variables

newTable <- synStore(Table(newSchema$properties$id, newData))
## 
Uploading [--------------------]0.00%   0.0bytes/265.0bytes  file158294a777a97     
Uploading [####################]100.00%   265.0bytes/265.0bytes (401.7bytes/s) file158294a777a97 Done...    
 [####################]100.00%   1/1   Done...

To download attached files in a table:

Set Annotations

A table schema is a Synapse entity. Annotations on table works the same way as annotations on any other entity types.

To set annotation on table, use synSetAnnotations() on the schema:

To view annotations on table, retrieve the schema:

Please visit synapser vignettes to read more about annotations.

Deleting Table

Deleting the schema deletes the whole table and all rows:

Queries

The query language is quite similar to SQL select statements, except that joins are not supported. The documentation for the Synapse API has lots of query examples.

For more details see the native reference documentation, e.g.:

synDelete(project)
## NULL
fileHandleIds <- lapply(files, function(f) file.remove(f$fileName))