2

I am doing a query (i.e. selecting multiple rows based on some condition) using PyTables, with the functions tables.Table.read() and tables.Table.read_where(). This is essentially based on numpy and pandas with NumExpr:

http://www.pytables.org/usersguide/tutorials.html http://www.pytables.org/cookbook/hints_for_sql_users.html https://github.com/PyTables/PyTables/blob/6782047b9223897fd59ff4967d71d7fdfb474f16/tables/table.py

In the "hints for sql users", the example for selecting multiple rows at once is this: 

rows = tbl.read_where('(sqrt(x**2 + y**2) <= 1) & (temperature < 100)')

Let's say I would prefer to do as query as follows: all rows that equal temperature 100 or equal temperature 90

rows = tbl.read_where('(temperature == 100) | (temperature == 90)')

This works perfectly. But I would like to pass through a list/array of "temperature values" to accomplish this task. 

temperatures = [80, 90, 100]
# reads in temperatures
# performs this query: 
rows = tbl.read_where('(temperature == 80) | (temperature == 90) | (temperature == 100)')

Is this possible? The idea would be I would write a function whereby users input a list of values to query, and it performs an OR query for each.

ShanZhengYang
  • 16,511
  • 49
  • 132
  • 234

1 Answers1

2

One possible solution is create expression by list comprehension:

temperatures = [80, 90, 100]

cond = '|'.join(['(temperature == ' + str(x) + ')' for x in temperatures])
print (cond)
(temperature == 80)|(temperature == 90)|(temperature == 100)
jezrael
  • 822,522
  • 95
  • 1,334
  • 1,252
  • Maybe works [this](http://pandas.pydata.org/pandas-docs/stable/indexing.html#special-use-of-the-operator-with-list-objects), I cannot test it. – jezrael Nov 03 '16 at 08:06